diff --git "a/69E2T4oBgHgl3EQfkweU/content/tmp_files/load_file.txt" "b/69E2T4oBgHgl3EQfkweU/content/tmp_files/load_file.txt" new file mode 100644--- /dev/null +++ "b/69E2T4oBgHgl3EQfkweU/content/tmp_files/load_file.txt" @@ -0,0 +1,1472 @@ +filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf,len=1471 +page_content='Exploring the Use of WebAssembly in HPC Mohak Chadha, Nils Krueger, Jophin John, Anshul Jindal, Michael Gerndt {firstname.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='lastname}@tum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='de Chair of Computer Architecture and Parallel Systems, Technische Universität München, Germany Shajulin Benedict shajulin@iiitkottayam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='ac.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='in Department of Computer Science and Engg.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Indian Institute of Information Technology Kottayam, Kerala Abstract Containerization approaches based on namespaces offered by the Linux kernel have seen an increasing popularity in the HPC community both as a means to isolate applications and as a format to package and distribute them.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, their adoption and usage in HPC systems faces several challenges.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These include difficulties in unprivileged running and build- ing of scientific application container images directly on HPC resources, increasing heterogeneity of HPC architectures, and access to specialized networking libraries available only on HPC systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These challenges of container-based HPC appli- cation development closely align with the several advantages that a new universal intermediate binary format called We- bAssembly (Wasm) has to offer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These include a lightweight userspace isolation mechanism and portability across oper- ating systems and processor architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In this paper, we explore the usage of Wasm as a distribution format for MPI- based HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, we present MPIWasm, a novel Wasm embedder for MPI-based HPC applications that enables high-performance execution of Wasm code, has low- overhead for MPI calls, and supports high-performance net- working interconnects present on HPC systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We evaluate the performance and overhead of MPIWasm on a production HPC system and AWS Graviton2 nodes using standardized HPC benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Results from our experiments demonstrate that MPIWasm delivers competitive native application per- formance across all scenarios.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, we observe that Wasm binaries are 139.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5x smaller on average as compared to the statically-linked binaries for the different standardized benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' CCS Concepts: • Software and its engineering → Process management.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Keywords: WebAssembly, Wasmer, Wasm, MPI, HPC Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Copyrights for components of this work owned by others than ACM must be honored.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Abstracting with credit is permitted.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Request permissions from permissions@acm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada © 2023 Association for Computing Machinery.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ACM ISBN 979-8-4007-0015-6/23/02.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='$15.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='00 https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/nnnnnnn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='nnnnnnn ACM Reference Format: Mohak Chadha, Nils Krueger, Jophin John, Anshul Jindal, Michael Gerndt and Shajulin Benedict.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2023.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Exploring the Use of We- bAssembly in HPC.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In The 28th ACM SIGPLAN Annual Sympo- sium on Principles and Practice of Parallel Programming (PPoPP ’23), February 25-March 1, 2023, Montreal, QC, Canada.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ACM, New York, NY, USA, 16 pages.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/nnnnnnn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' nnnnnnn 1 Introduction Linux containers, due to their portability and high availability, have become the de-facto standard for developing, testing, and deploying a wide range of applications from enterprise to web services in cloud environments [29].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is because containers enable users to package their application along with its custom software dependencies as a single unit into easy-to-deploy images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Motivated by their popularity in the cloud, containers have also seen a growing interest in the HPC community [30, 74, 90].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For HPC systems, containers provide flexibility to users and allow them to define custom software stacks, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', user-defined software stack (UDSS) for their large- scale scientific applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, they enable easy, reli- able, and verifiable environments that can be reproduced in the future.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, several HPC-focused containerization solutions, such as Charliecloud [72], Shifter [49], Singular- ity [58], Podman [48], and Sarus [33] have been introduced.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In contrast to previous approaches, this paper investigates us- ing a new novel technology called WebAssembly (Wasm) [52], dubbed as an alternative to Linux containers [81], for packag- ing and distributing HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Despite their increasing popularity, the adoption and us- age of containers in HPC systems is still significantly lim- ited [36].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This can be attributed to the several challenges commonly faced by users in running and building container images for their applications on HPC systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For execut- ing containers, most containerization solutions require root privileges which are not possible for normal HPC users due to shared filesystems and their UNIX permissions in HPC.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While HPC-focused containerization solutions such as Sin- gularity [58] and Podman [48] support rootless-containers through fakeroot [61], their current implementations do not support distributed filesystems such as GPFS commonly found on HPC systems [70, 80].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, as argued by [71], building Open Container Initiative (OCI) [69] compliant con- tainer images on HPC resources by unprivileged (normal) arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='03982v1 [cs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='DC] 10 Jan 2023 PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' users where the applications will eventually run is signifi- cantly hard and requires support from the supercomputing center.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is because most container building solutions such as Docker [43] also require root privileges.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, most users use their own local systems for building/developing their application container images and then transfer the built image to a login/front-end node of an HPC system for execution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, this scenario leads to several problems in container- based HPC application development.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' First, HPC nodes are becoming more heterogeneous [11] with different processor architectures such as x86_64 or aarch64 and have special- ized accelerators such as GPUs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As application performance is critical in HPC, compiling an application using the specific microarchitectural features of a particular processor is signifi- cantly important.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While building container images for mul- tiple platforms either by cross-compiling HPC applications or by emulation with QEMU is possible with plugins such as build-x [44], it is not widely supported by HPC application build procedures and requires the presence of specific Linux kernel features (binfmt_misc [62]).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, testing and developing HPC applications offers insights only on the target system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In addition, most container images can range from several MiBs to several GiBs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, frequent network transfers from the local to the HPC system can be cumber- some.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Second, building HPC applications requires access to specialized networking libraries and licenses to compilers that are not available on the local user systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Finally, while different containerization solutions have almost no impact on the performance of the containerized application [72, 75, 84], building high-performant HPC application container images is non-trivial, involves a steep learning curve, and requires knowledge about specific MPI library versions (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Open- MPI [10] 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0) and high performance network interconnect hardware (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Intel OmniPath [4]) and libraries (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Intel Performance Scaled Messaging [5]) present on the target sys- tem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These challenges of container-based HPC application development closely align with the several advantages and core problems that Wasm [52] aims to solve.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm is a low-level, statically typed universal binary in- struction format for memory-safe, sandboxed execution in a virtual machine.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It offers portability across modern proces- sor architectures and operating systems, fast execution, and a low-level memory model [52].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Although originally meant for execution in Web browsers, due to its simplicity and gen- erality, Wasm has seen widespread adoption and usage in non-Web domains such as serverless computing [78], edge computing [47, 53], and Internet of Things [51].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It does not require garbage collection and is designed to be a universal compilation target with mature support for programming lan- guages with an LLVM [59] front-end such as C, C++, C#, and Rust [32, 35, 40, 45, 77].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 1 demonstrates a general workflow for using Wasm in HPC.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Developers can compile their HPC applications to Wasm once on their local systems ahead-of-time (AoT) and HPC Application WebAssembly x86_64 aarch64 WebAssembly embedder Compile Figure 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' An HPC application can be compiled to WebAssem- bly and distributed to multiple platforms where it can be executed efficiently by a supporting WebAssembly embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' distribute it across multiple platforms instead of distribut- ing source code or building application containers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Typically, Wasm binaries have a smaller size as compared to native x86_64 binaries [52, 56, 91].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, the resulting binary can be executed on any platform using a standalone Wasm embedder [52].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm embedder serves two major purposes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' First, it provides an isolated execution environ- ment for running a Wasm binary on a platform.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In contrast to container-based approaches that utilize different Linux names- paces [71] for isolation and security, Wasm provides light- weight isolation at the application level based on software fault isolation (SFI) [85] and control flow integrity (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Second, it is responsible for compiling Wasm binaries to native machine code, either by using Just-in-Time (JIT) en- gines at the time of execution, or AoT by using the same JIT engines or AoT compilers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Note that, Wasm binaries can be executed by normal users and are completely unpriv- ileged.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Several open-source standalone embedders such as Wasmer [87], Wasmtime [37], and Wasm3 [79] are currently available.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, none of them support the execution of HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As the first step towards bringing Wasm to the HPC ecosys- tem, we only focus on MPI-based [6] HPC applications in this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We chose MPI due to its understanding and in- fluence in the HPC community [34].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, our key contributions are: We implement and present MPIWasm, a novel Wasm embedder for MPI-based HPC applications based on Wasmer [87].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPIWasm enables high performance exe- cution of Wasm code, has low-overhead for MPI calls through zero-copy memory operations, and supports high-performance networking interconnects such as Intel OmniPath [4].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We demonstrate with extensive experiments the low- overhead and performance of MPIWasm using standard- ized HPC benchmarks on a production HPC system and AWS Graviton2 [1] nodes based on the x86_64 and the aarch64 architectures respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We elaborate on the different possible future directions for using Wasm in the HPC ecosystem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WA口口Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada The rest of this paper is structured as follows.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' §2 provides a detailed overview on Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In §3, we describe our embedder MPIWasm in detail.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our experimental results are presented in §4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In §5, we describe the different possible directions for using Wasm in the HPC ecosystem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' §6 describes some previous approaches related to our work.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In §7, we conclude the paper and present an outlook.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2 A primer on WebAssembly 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 WebAssembly Overview WebAssembly (Wasm) was introduced in 2015 as an alter- native to JavaScript for web-browser based applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It superseded asm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='js [67], a previous attempt by Mozilla which focused on a subset of Javascript code that can be optimized AoT.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' When an application is compiled to Wasm, the resulting binary is called a module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm modules contain function definitions, declarations of global variables, tables, and a lin- ear memory address space.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' All of the application code in Wasm is organized in functions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The conceptual machine in Wasm is stack-based and does not contain registers, there- fore all instructions pop their operands from the stack of the machine.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, since application control flow is an explicit part of the module and Wasm operations are typed, it is possible to statically predict the layout of the stack at any point in the program which allows compilers to trans- late the stack semantics to a register-based instruction set.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Similar to other higher-level programming languages, Wasm allows the definition of global variables that are not scoped to a specific function or block.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Tables in Wasm modules are used for storing references to functions [52].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm ISA currently supports only four data types for variables: (i) i32, 32-bit integers, (ii) i64, 64-bit integers, (iii) f32, 32-bit IEEE 754 floating point numbers, and (iv) f64 64-bit IEEE 754 floating point numbers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For constructing, complex types a combination of these basic types is commonly used.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm provides the capability for data and code to be shared between the module and its embedder using the import/export system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' All of the function definitions that can occur in a Wasm module can be imported from the embedder instead of being defined within it.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Similarly, function definitions that are present in the module can be exported so that the embedder can utilize them (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 WebAssembly Security and Sandboxing Model Wasm utilizes software fault isolation techniques (SFI) [85] to sandbox the executing Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' By default, a Wasm module cannot interact with the host system or perform I/O operations of any kind.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Any system interaction that is to be initiated by the Wasm module’s code must be done through the functions imported from the embedder (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, the embedder can act both as a translation layer and as an arbiter to enforce isolation requirements.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a translator, it 1 (type (;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=') (func (param i32) (result i32))) 2 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3 (type (;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=') (func (param i32 i32) (result i32))) 4 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 5 (type (;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='14;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=') (func (param i32 i32 i32 i32 i32 i32) 6 (result i32))) 7 (type (;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='15;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=') (func (param i32 i32 i32 i32) (result i32 ))) 8 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 9 (import "wasi_snapshot_preview1" "path_open" 10 (func $__wasi_path_open (type 22))) 11 (import "wasi_snapshot_preview1" "fd_close" 12 (func $__wasi_fd_close (type 1))) 13 (import "wasi_snapshot_preview1" "fd_seek" 14 (func $__wasi_fd_seek (type 23))) 15 (import "wasi_snapshot_preview1" "fd_read" 16 (func $__wasi_fd_read (type 15))) 17 (import "wasi_snapshot_preview1" "proc_exit" 18 (func $__wasi_proc_exit (type 0))) 19 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 (export "_start" (func $_start )) 21 (export "memory" (memory 0)) Listing 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Example representation of a compiled C++ appli- cation’s Wasm module using the WASI-SDK in WebAssembly text format (WAT) [68].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Ellipses signify sections that are omitted for brevity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' is possible for the embedder to provide a common interface to the Wasm module even though the underlying system may have different native interfaces, while as an arbiter it is possi- ble for the embedder to restrict access of the Wasm module to system resources based on an application-level security policy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For instance, it is possible for the embedder to allow file I/O only to files that reside in a specific directory to iso- late the Wasm module from the rest of the filesystem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While in principle similar to kernel-level system call filtering tech- niques such as Seccomp-BPF [83] on Linux, performing such filtering on the application level allows to define semantically more meaningful policies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Wasm, all memory access is confined to a module’s lin- ear memory which is separate from the code space.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Currently, the Wasm specification [52] supports 32-bit addresses to in- dex the memory that a module has access to.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While this limits a single module’s memory to 4GiB, it also enables hardware accelerated bound checks of memory accesses at runtime [42].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' If an embedder is a process with a 64-bit memory address space, it can safely execute an untrusted Wasm module in its memory space without requiring additional isolation by re- serving a continuous range of virtual memory for the module to use.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Not all pages in this range need to be mapped to phys- ical memory, it is sufficient to only map the required number of pages to fit the amount of memory used by the module at a given point in time.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This ensures that a Wasm module can only operate in its own execution environment and cannot corrupt the memory of the embedder, since any out-of-bounds memory access will result in a page fault which can then be handled by it.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, since the memory instructions in Wasm’s specification [52] work with offsets, it is not possible to read and write to arbitrary memory locations in Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In the assembly produced by C programs, where a func- tion call is expressed as a jump instruction to the address of the function’s first instruction, a typical exploit is to change this address to take control of the program’s control flow.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 typedef int MPI_Comm;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2 typedef int MPI_Datatype;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4 int MPI_Init(int* argc , char *** argv);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 5 int MPI_Finalize(void);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 6 int MPI_Send( 7 const void* buf , int count , MPI_Datatype datatype , 8 int dest , int tag , MPI_Comm comm 9 );' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 10 int MPI_Recv( 11 void* buf , int count , MPI_Datatype datatype , 12 int source , int tag , MPI_Comm comm , MPI_Status* status 13 );' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Excerpt of the custom MPIWasm mpi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='h header file.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, such exploits are not possible with Wasm since it features control flow integrity by enforcing structured pro- gram control flow.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is because of two reasons.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' First, in Wasm, a function is represented as an index in a table (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1) which adds an additional level of indirection to express the function address.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Second, the Wasm specification prevents constructing arbitrary memory addresses [42] and the sepa- ration of the embedder and the module’s memory prevents overwriting function instructions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3 WebAssembly System Interface Since Wasm was originally designed for web browsers, a sys- tem interface that targets POSIX environments and enables execution of Wasm modules on them was not part of the orig- inal specification [52].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To overcome this, the WebAssembly System Interface (WASI) specification [89] was designed.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WASI specifies the interface an embedder needs to implement to execute most POSIX applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Embedders that imple- ment the WASI specification will be able to run any generic application compiled with the WASI-SDK [28].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The WASI-SDK includes the clang compiler and its own C library based on musl libc that call WASI systemcalls imported from the em- bedder instead of relying on Linux systemcalls [22].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Note that, due to the ubiquity of glibc [26] on Linux systems, some ap- plications have come to depend on glibc-specific functions or behavior.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Such applications will require modifications before they can be compiled to a WASI-compliant Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 1 shows a compiled Wasm module of a C++ appli- cation using the WASI-SDK in the WebAssembly text format (WAT).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WAT is a human readable format that enables de- velopers to examine the source code of a Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It can be observed that the module contains several functions with integers as parameter and return types (Lines 1-7) (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1), imports WASI functions (Lines 9-18), and exports its _start (main function) and memory (Lines 20-21).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Exporting these two definitions allows the embedder that executes this module to call its entrypoint function and to read from and write to the module’s linear memory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While the import statements on Lines 9-16 enable the Wasm module to open and read from a file, the function proc_exit is used by the embedder to handle the termination of the application, e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', by deallocat- ing the memory reserved for the module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the module to 1 (import "env" "MPI_Init" ( 2 func $MPI_Init (param i32 i32) (result i32) 3 )) 4 (import "env" "MPI_Finalize" (func $MPI_Finalize (result i32 ))) 5 (import "env" "MPI_Send" ( 6 func $MPI_Send (param i32 i32 i32 i32 i32 i32) (result i32) 7 )) 8 (import "env" "MPI_Recv" ( 9 func $MPI_Recv (param i32 i32 i32 i32 i32 i32 i32) 10 (result i32) 11 )) Listing 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WAT representation of module imports that corre- spond to the functions shown in Listing 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' execute, the imported functions need to be implemented by the embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3 MPIWasm In this section, we describe MPIWasm, our embedder for executing Wasm modules that utilize functions from the MPI standard in detail.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 Overview The purpose of MPIWasm is to support the execution of MPI applications compiled to Wasm on HPC systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To facili- tate its adoption and suitability in HPC environments, it (i) supports high-performance execution of MPI-based HPC ap- plications compiled to Wasm (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3), (ii) has low-overhead for MPI calls through zero-copy memory operations (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6), and (iii) supports high-performance interconnects such as Infiniband [64] and Intel OmniPath [4].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These network inter- connects are utilized by MPI libraries on HPC systems for high-performance inter-rank communication.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To enable the immediate support for network interconnects present on mod- ern HPC systems, MPIWasm links against the MPI library on the target HPC system at runtime and provides a translation layer between the Wasm module and the host1 MPI library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, the developer doesn’t need to be aware about the particular networking libraries or network interconnects present on the target HPC system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Depending on the partic- ular host MPI library such as OpenMPI [10] or MPICH [8], MPIWasm needs to be built separately.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Both of these libraries are currently supported by MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our embedder currently supports the execution of MPI applications written in C/C++ and conforming to the MPI-2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 standard [65].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Integrating the support for MPI-3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 [66] is of our interest for the future but is out of scope for this work.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We chose to focus on C/C++ applications due to the stability and maturity of the Wasm backend in the LLVM/Clang [59] project since llvm-8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As the base for MPIWasm, we use the open-source Wasm embedder called Wasmer [87].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasmer supports the execution of Wasm modules on three major plat- forms, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Linux, Windows, and macOS, and supports both x86_64 and aarch64 instruction set architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, it implements the WASI specification (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3) and provides 1We use the term target and host interchangeably for the system on which the Wasm module is executing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada ergonomic mechanisms to define additional functions that are provided to the module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This dynamic extension of the embedder’s functionality enables the addition of MPI func- tions to the functionality it provides to the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For implementing MPIWasm, we use the Rust programming lan- guage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is because of two reasons.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' First, it provides high performance comparable to C/C++ with memory-safety [82].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Second, it has extensive support and documentation for em- bedding Wasmer and using it as a library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 Compiling C/C++ MPI applications to Wasm Most MPI applications expect POSIX functionality to be available in their execution environment, for instance the abil- ity to read from and write to file descriptors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WASI (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3) defines the WebAssembly exports that enable Wasm mod- ules that target it to call most of the functions defined in the C standard libraries shipped on POSIX systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, the WASI-SDK [28] combines the clang compiler and the wasi-libc C library to enable the compilation of C/C++ applications that only make use of POSIX functions and no additional libraries to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The compilation of C/C++ MPI applications is not supported by the stock WASI-SDK.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, we implement a custom mpi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='h MPI header file and add it to the WASI-SDK.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The header file includes the definitions for the different MPI types such as MPI_Op, MPI_Comm, and MPI_Datatype and the definition for the MPI_Status struc- ture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, it defines the signatures for the MPI functions according to the MPI-2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 [65] standard.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' An excerpt from the header file is shown in Listing 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It is a reduced version of a traditional header file found with MPI libraries with most types defined as integers (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' By combining our header file with the WASI-SDK, a C/C++ MPI application conforming to the MPI-2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 standard can be compiled to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, to facilitate the ease-of-use and enable adoption, we imple- ment a custom python-based tool that simplifies the entire compilation process for MPI applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 3 shows the different MPI-specific imports present in a Wasm module corresponding to the functions shown in Listing 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPIWasm provides definitions for these imports to enable the execution of MPI-based HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In addition, it supports the WASI specification which enables the POSIX functionality for MPI applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3 Executing Wasm Code with High Performance There exist several strategies for executing Wasm modules.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These include using an interpreter [79], Ahead-of-Time (AoT) compilation [37], or Just-in-Time (JIT) compilation [37].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, for HPC systems the most useful approach is trans- lating the Wasm instructions (Wasm ISA) to the native instruc- tion set of the host machine before the application is executed, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', AoT.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, MPIwasm builds on the code genera- tion infrastructure provided by Wasmer [87].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasmer currently supports three compiler backends, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', Singlepass [86], Cranelift [3], and LLVM [59].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The SinglePass compiler Table 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Comparing compile duration and performance for the different compiler backends supported by Wasmer [87] for the HPCG [73] Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm module was generated using our WASI-SDK (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm module is executed using MPIWasm on an x86_64 system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Compiler Compile Duration (ms) Single-Core Performance (GFLOP/s) Singlepass [86] 52 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3769 Cranelift [3] 150 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3240 LLVM [59] 2811 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5426 is designed to emit machine code in linear time and does not perform many code optimizations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Cranelift com- piler is completely based on Rust and is similar to LLVM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With Cranelift, the WASM instructions are first translated to the intermediate representation (IR) of Cranelift, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', (Cranelift-IR) which are then translated to the native in- struction set of the host machine by taking microarchitecture- specific optimizations into account.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the other hand, with LLVM the Wasm ISA is first translated to LLVM-IR followed by the generation of native machine code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Cranelift-IR is similar to LLVM-IR but at a lower level of abstraction which hinders mid-level code optimizations2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' At the end of the com- pilation process, all three compilers produce a shared ob- ject, which can be loaded with a fast dlopen call using the libloading library [27].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Table 1 shows a comparison of the compile-time and run- time performance of the three different compilers supported by Wasmer for the HPCG benchmark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While LLVM is the slow- est to compile the Wasm module, it also results in the fastest runtime performance for the HPCG application.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, we chose LLVM as the compiler backend in MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To offset the longer compilation times required by LLVM as com- pared to the other two compilers, we implement a caching mechanism for the generated machine code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our caching mechanism builds on the FileSystemCache [46] provided by Wasmer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In our implementation, we generate a hash for each Wasm module using the Blake-3 hash function [2].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, we store the generated shared object from LLVM as the generated hash in the local filesystem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, any changes to the Wasm module lead to the generation of a new hash which triggers the recompilation of the module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, repeated execution of the same application on a system with MPIWasm will not lead to recompilation overhead for execution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 Filesystem Isolation with MPIWasm Since in Wasm all system interactions by the application have to be performed by calling functions implemented by the embedder (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2,§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3), it enables the embedder to place addi- tional restrictions on their use and to employ checks on the arguments supplied to them.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Wasmer, all exported functions 2A more detailed discussion between Cranelift-IR and LLVM-IR can be found here [25].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Memory address space of MPIWasm with an instan- tiated Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' All memory access instructions to the Wasm module’s linear address space are given offsets relative to the base address.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' that handle file I/O perform their own permission handling that is separate from the one employed by the OS.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This in- process indirection of filesystem accesses allows Wasmer to present a virtual directory tree to the Wasm module that only contains directories that the module is allowed to access.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In addition, access rights to individual directories can be more restrictive than the permissions granted to the user that is exe- cuting the embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For instance, a user can have read and write access to their home directory and all of its subdirecto- ries, but grant read-only access to one specific subdirectory to a Wasm module executed by the embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPIWasm ex- poses this isolation functionality with its -d flag that grants read-write access to the given directory to the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Note that the full absolute path to the exposed directories is not presented to the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In the virtual direc- tory tree presented to it, all of the subdirectories it has been given access to are direct children of the root directory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This approach to mapping directory paths avoids exposure of in- formation contained in the full path to the directories, such as a username in the case of a home directory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 Translating from Wasm to Host Memory Address A major part of the Wasm security model is the separation of the host and the module’s linear memory address space (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Since it is the responsibility of the Wasm embedder to uphold capability restrictions, protecting it’s data structures from unintended or malicious access by the modules’ code is significantly important.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, this separation presents a challenge for supporting MPI applications, because the MPI API is based on the library being able to read and write di- rectly to the memory of the application.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The executing Wasm- based MPI application can only provide memory addresses in its own linear memory address space, while the target MPI library requires addresses in the host memory address space.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For executing Wasm modules, MPIWasm reserves a part of its own address space for use by the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, every byte contained in this range can be addressed either with a memory address in the module’s memory space or with a memory address in the embedder’s (host’s) memory space.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, while instantiating the module’s linear mem- ory, MPIWasm records its base address.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, it is possible to convert an address from the linear address space of the Wasm module to the embedder’s address space and vice- versa by treating the address in the linear address space as an offset relative to the module’s base address.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is shown in Figure 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In particular, MPIWasm directly converts 32-bit Wasm pointers that refer to the module’s linear address space to 64-bit pointers that refer to the embedder’s address space and vice-versa.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, MPIWasm directly utilizes the MPI library present on the host system without copying any data from the module’s address space to a different location, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', it supports zero-copy memory operations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our mechanism for memory address translation does not violate memory-safety because: (i) a malicious Wasm module cannot violate control flow integrity (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2) and (ii) since the size of the linear memory is always known, MPIWasm can perform runtime bound checks for all memory accesses.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, a module cannot access the memory of the embedder or the memory of the underlying operating system unless explicitly given access to it.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 Translating MPI Datatypes MPI is implemented as a library with the most common being OpenMPI [10], MPICH [8], and MVAPICH [9].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Hence, it does not guarantee an Application Binary Interface (ABI) and interoperability between libraries.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This means that chang- ing the MPI implementation requires recompilation of the entire application code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' One of the reasons for ABI incom- patibility is that the MPI standard does not specify explicit types for its datatypes such as MPI_Op and their implemen- tation is completely up to the MPI library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, since Wasm modules are designed to be portable not just between the different MPI libraries but also between different CPU architectures, it becomes necessary to add an abstraction be- tween the datatypes used by the host’s MPI library and the datatypes exposed to the Wasm module by MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' An abstraction is possible since most MPI datatypes such as MPI_Comm, MPI_Datatype and MPI_Op are opaque to the ap- plication and only used as arguments to MPI functions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI- Wasm defines most MPI datatypes as 32-bit integers from the perspective of the Wasm module (Listing 2) and transpar- ently translates these datatypes to the host equivalents (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We use integers as datatypes since MPIWasm internally uses IDs to identify data structures that it creates on behalf of the module in order to communicate with the host MPI library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7 Implementing MPI Functions in MPIWasm Wasm imports are referred to by namespace and name of the definition to import.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' By default, any symbols that are not defined while compiling C/C++ applications to Wasm will be resolved by making them imports of the module in the env namespace.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is also demonstrated in Listing 3 with the function imports related to the MPI standard.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI- Wasm provides definitions for all these functions with the same name as the original MPI function and exports them in the env namespace.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For implementing these functions, we 0x0 OxFFFF_FFFFExploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada combine the memory address and MPI datatype translations as described in §3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 and §3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, we maintain a structure called Env that stores the global state required by these translations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This structure includes infor- mation about the memory allocated to the Wasm module, it’s base pointer (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5) and information about the different used datatypes such as MPI_Comm by the module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For directly utiliz- ing the host MPI library, we use the project rsmpi [7] in MPI- Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' rsmpi provides MPI bindings for Rust and supports OpenMPI [10] and MPICH [8].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It utilizes the rust-bindgen project to generate foreign function interfaces tailored to spe- cific MPI libraries.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Each MPI function in MPIWasm directly calls the equivalent function in rsmpi with the appropriate arguments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While for most functions in the MPI-2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 standard MPI- Wasm directly defers the execution to the host MPI library, the implementation of the MPI functions MPI_Alloc_mem and MPI_Free_mem is done differently.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With these functions, it is possible to allocate memory for use with other MPI functions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' When a Wasm module calls MPI_Alloc_mem, it expects a 32- bit memory address in the module’s address space, while call- ing the MPI_Alloc_mem function of the host MPI library re- turns a 64-bit memory address in the embedder’s memory ad- dress space which is not inside the chunk of memory reserved for the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To overcome this, MPIWasm only supports MPI_Alloc_mem and MPI_Free_mem if the Wasm module defines and exports the functions malloc and free.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' When MPI_Alloc_mem is called, MPIWasm simply invokes the exported malloc and receives a suitable 32-bit module memory address.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This address can then be used as the return value for MPI_Alloc_mem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We implement MPI_Free_mem in a similar way.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='8 Limitations The Wasm specification currently assumes little-endian byte order for multi-byte values [52] in execution environments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' By giving direct access to the Wasm module’s memory to the host MPI library, we assume that the byte order of values in the module address space and embedder address space is the same.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, MPIWasm does not support big-endian CPU architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is not a disadvantage since most processor architectures in HPC systems are little-endian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, due to the current linear 32-bit memory space for a Wasm module, HPC applications compiled to Wasm cannot have more than 4GiB of memory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The support for 64-bit memory addresses is an important milestone for the Wasm specification and is highlighted in the Wasm Memory64 proposal [20], but is out of scope for this work.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4 Experimental Results In this section, we present performance results for our embed- der MPIWasm across different processor architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For 1 mpirun -np .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='/ mpiWasm mpi -app.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='wasm Listing 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Executing MPI applications compiled to Wasm with MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Table 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Comparing the size of native dynamically-linked, statically-linked, and Wasm binaries for the different MPI applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The native applications are compiled for the x86_64 architecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Application Native Size Dynamic (KiB) Native Size Static (MiB) Wasm Size (KiB) Intel MPI Benchmarks [55].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1087 27 893 HPCG [73].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 164 26 722 IOR [60].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 364 16 315.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='32 IS [31].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 36 15 57.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='88 DT [31].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 40 15 49.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='51 all our experiments, we follow best practices while reporting results [54].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 System Description For analyzing the performance of our implemented Wasm embedder, we use two systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' First, a production HPC clus- ter located at our institute, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', SuperMUC-NG.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Second, an AWS EC2 virtual machine (VM) instance with the Gravi- ton2 processor [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our HPC cluster contains eight islands comprising a total of 6480 compute nodes based on the Intel Skylake-SP architecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Each compute node has two sockets, comprising two Intel Xeon Platinum 8174 processors, with 24 cores each and a total of 96GiB of main memory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The nominal operating core frequency for each core is 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='10 GHz.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Hyper-Threading and Turbo Boost are disabled on the system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The internal interconnect on our system is a fast Intel Omni- Path [4] network with a bandwidth of 100 Gbit/s.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, our cluster provides a general parallel filesystem based on the Lenovo DSS-G for IBM Spectrum Scale [19] with an aggregate bandwidth of 200 GiB/s.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For our experiments, we use up to 128 nodes of the HPC system, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', 6144 cores.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the other hand, the AWS Graviton2 processor based on the 64-bit ARMv8-A Neoverse-N1 [24] architecture consists of 32 cores each with a nominal frequency of 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='50 GHz and a total main memory of 64GiB.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We limit our experiments to one node for the Graviton2 processor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 HPC Benchmarks For our experiments with MPIWasm, we use the Intel MPI Benchmarks [55], two benchmarks from the the NASA Ad- vanced Supercomputing (NAS) Parallel Benchmark (NPB) suite [31], the IOR benchmark [60], and the High Perfor- mance Compute Gradient (HPCG) benchmark [73].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Intel MPI benchmarks perform a set of MPI perfor- mance measurements for point-to-point and global communi- cation operations for a range of message sizes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We use them since they characterize the performance of a cluster and are an indication of the efficiency of the used MPI implementa- tion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The NPB suite includes a set of benchmarks that aim to evaluate the overall performance of HPC clusters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Due to the support for compiling Fortran to Wasm being in the early stages (§5), only the Integer Sort (IS) and Data Transfer (DT) PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 1 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 Bytes Iteration Time (usec) PingPong (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 101 102 Bytes Iteration Time (usec) PingPong (time) > 1024 Bytes Native WASM (a) PingPong.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 2 4 Bytes Iteration Time (usec) Sendrecv 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 101 102 103 104 Bytes Iteration Time (usec) Sendrecv 6144 Ranks (time) > 1024 Bytes Native WASM (b) SendRecv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0 20 40 Bytes Iteration Time (usec) Bcast 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 100 102 104 Bytes Iteration Time (usec) Bcast 6144 Ranks (time) > 1024 Bytes Native WASM (c) Broadcast.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 22 24 26 28 210 40 60 Bytes Iteration Time (usec) Allreduce 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 102 104 Bytes Iteration Time (usec) Allreduce 6144 Ranks (time) > 1024 Bytes Native WASM (d) AllReduce.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0 1 2 104 Bytes Iteration Time (usec) Allgather 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 217 105 106 Bytes Iteration Time (usec) Allgather 6144 Ranks (time) > 1024 Bytes Native WASM (e) AllGather.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0 1 2 105 Bytes Iteration Time (usec) Alltoall 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 106 Bytes Iteration Time (usec) Alltoall 6144 Ranks (time) > 1024 Bytes Native WASM (f) Alltoall.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 22 24 26 28 210 0 10 20 30 Bytes Iteration Time (usec) Reduce 768 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 100 102 104 Bytes Iteration Time (usec) Reduce 768 Ranks (time) > 1024 Bytes Native WASM 22 24 26 28 210 0 20 40 60 Bytes Iteration Time (usec) Reduce 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 100 102 104 Bytes Iteration Time (usec) Reduce 6144 Ranks (time) > 1024 Bytes Native WASM (g) Reduce.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='20 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='22 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='24 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='26 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='28 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='210 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='50 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='100 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='150 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Iteration Time (usec) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Gather 768 Ranks (time) ≤ 1024 Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Native ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='WASM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='212 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='214 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='216217218219220 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='100 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='102 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='104 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Iteration Time (usec) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Gather 768 Ranks (time) > 1024 Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Native ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='WASM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='20 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='22 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='24 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='26 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='28 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='210 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='200 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='400 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='600 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Iteration Time (usec) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Gather 6144 Ranks (time) �� 1024 Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Native ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='WASM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='212 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='214 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='216 217 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='100 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='102 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='104 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Iteration Time (usec) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Gather 6144 Ranks (time) > 1024 Bytes ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Native ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='WASM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='(h) Gather.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0 100 200 300 Bytes Iteration Time (usec) Scatter 768 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216217218219220 101 103 105 Bytes Iteration Time (usec) Scatter 768 Ranks (time) > 1024 Bytes Native WASM 20 22 24 26 28 210 0 1,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='000 2,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='000 Bytes Iteration Time (usec) Scatter 6144 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 217 101 103 105 Bytes Iteration Time (usec) Scatter 6144 Ranks (time) > 1024 Bytes Native WASM (i) Scatter.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Performance comparison of the Intel MPI benchmarks for MPIWasm and their native execution on our HPC system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='8 1 Bytes Iteration Time (usec) PingPong (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 100 101 102 Bytes Iteration Time (usec) PingPong (time) > 1024 Bytes Native WASM (a) PingPong.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 1 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 Bytes Iteration Time (usec) Sendrecv 32 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 100 101 102 103 Bytes Iteration Time (usec) Sendrecv 32 Ranks (time) > 1024 Bytes Native WASM (b) SendRecv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 22 24 26 28 210 2 4 6 8 Bytes Iteration Time (usec) Allreduce 32 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 101 102 103 104 Bytes Iteration Time (usec) Allreduce 32 Ranks (time) > 1024 Bytes Native WASM (c) AllReduce.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 0 10 20 Bytes Iteration Time (usec) Allgather 32 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 102 103 104 105 Bytes Iteration Time (usec) Allgather 32 Ranks (time) > 1024 Bytes Native WASM (d) AllGather.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 20 40 Bytes Iteration Time (usec) Alltoall 32 Ranks (time) ≤ 1024 Bytes Native WASM 212 214 216 218 220 222 102 104 Bytes Iteration Time (usec) Alltoall 32 Ranks (time) > 1024 Bytes Native WASM (e) Alltoall.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 12 4 8 16 32 0 10 20 Ranks GFLOP/s HPCG GFLOPS Native WASM 12 4 8 16 32 0 50 100 150 Ranks GB/s HPCG Bandwidth Native WASM (f) HPCG Figure 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Performance comparison of selected Intel MPI benchmarks and HPCG for MPIWasm against their native execution on the AWS Graviton2 Processor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' benchmarks from this suite were used since they are written in pure C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The IS benchmark performs bucketed parallel sort- ing of integers across all participating processes, while the DT benchmark tests the communication and the performance of 64-bit floating point operations of a HPC cluster by send- ing data through a topology of nodes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We use the topologies Black-Hole (bh), White-Hole (wh), and Shuffle (sh) for the DT benchmark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For our experiments, we use the classes C and B for the IS and DT benchmarks respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The IOR Benchmark measures the filesystem I/O performance avail- able to MPI processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It supports multiple backends that utilize different APIs to perform system I/O.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For our experi- ments with MPIWasm, we use the POSIX API backend since the POSIX filesystem APIs are included in the WASI specifi- cation (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3, §3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The HPCG benchmark aims to evaluate the real-world performance of HPC systems by solving a sys- tem of linear equations with the conjugate gradient method.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For our experiments, we use the default available problem Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada size for HPCG.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Note that, in our experiments we use the ver- sions 2019 Update 6 and 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 for the Intel MPI and NAS parallel benchmarks respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3 Experiment Setup For all our experiments, we execute the benchmarks in a pure- MPI configuration without shared memory parallelization with OpenMP, as it is currently not supported by MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We use OpenMPI-4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0 as the MPI library since it is available on our HPC system and can be easily installed on the AWS Graviton2 nodes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For compiling the native applications on our HPC system (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1), we use the clang-11 compiler, while for the AWS Graviton2 node, we use the gcc7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 compiler.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In both cases, the applications were compiled with the -O3 optimization flag.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For compiling the different benchmarks to Wasm, we use the clang-11 compiler along with our cus- tomized WASI-SDK with -O3 -msimd128 flags for both test systems (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The -msimd128 flag enables the generation of SIMD instructions in Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We compile the applications to Wasm only once on our local systems and execute them directly with MPIWasm on the different test systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We build MPIWasm on our local system for the different plat- forms, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', x86_64 and aarch64 with OpenMPI to generate bindings for rsmpi (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, we directly exe- cute the applications compiled to Wasm on the test systems as shown in Listing 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Each MPI rank corresponds to one instance of the embedder with it’s own Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The native applications were executed directly using mpirun.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 Comparing Wasm Binary Size Table 2 shows the comparison between the absolute binary sizes for the different applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The static versions of the binaries are generated by supplying the -static flag to the clang-11 compiler (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3) and linking the different applica- tions with the static versions of the required libraries such as libmpi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='a, libopen-rte.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='a, and libz.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To this end, we made necessary changes to the Make [50] and CMake [39] files used by the different applications (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While the stack- based instruction set and compact binary format give Wasm the potential to produce smaller binaries for the same appli- cations as compared to the native dynamically-linked bina- ries [52], three out of five applications that we used had a bigger binary size when compiled to WebAssembly in com- parison to the equivalent dynamically-linked native binary.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While Wasm can benefit from a smaller representation on a function-by-function basis, in practice dynamically-linked native binaries can offset that advantage by being able to rely on commonly used libraries to be present on the system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For instance, a native binary can dynamically link against glibc, while a Wasm binary must statically include functions from wasi-libc (§2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, in contrast to containers, Wasm binaries are significantly smaller making them more feasible for application distribution in HPC environments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In addi- tion, Wasm binaries are 139.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5x smaller on average than the statically-linked binaries of the different applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This is because the linker, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', lld copies all library routines from the different libraries used by an application into the binary during static linking.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 Benchmarking MPIWasm Figure 3 and Figure 4 show the iteration times for the different Intel MPI benchmarks for their native execution as compared to their execution with MPIWasm on our HPC system and the AWS Graviton2 processor respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For execution with MPIWasm, the iteration times don’t include the time required for compiling the Wasm modules to native machine code (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To avoid repetition, we omit some results for the Graviton2 processor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Error bars in the graphs represent mini- mum and maximum values for iteration timings as reported by the Intel MPI Benchmarks, while points in the graphs represent the average timings as reported by the benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the PingPong benchmark using MPIWasm leads to a geometric mean (GM) average slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='05x for the x86_64 system and a GM average speedup of 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='01x for the aarch64 system across all message sizes (Figures 3a, 4a).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We calculate this value by dividing the metric t_avg_us reported by the benchmarks for their native execution by the value reported for execution with MPIWasm, followed by a GM of the obtained values.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For computing slowdown, we subtract one from the obtained GM value.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We observe a maximum bandwidth of 12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='80 GiB/s and 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='98 GiB/s for the native exe- cution of the PingPong benchmark on the HPC and Graviton2 processor respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the other hand, with MPIWasm, we observe a maximum bandwidth of 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='44 GiB/s and 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='61 GiB/s on the two systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the SendRecv benchmark, we observe a GM average slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='06x and 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='07x with MPIWasm across all message sizes on the x86_64 and aarch64 systems respectively (Figures 3b, 4b).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the native version of the benchmark, we observe a maximum bandwidth of 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='24 GiB/s and 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='01 GiB/s on the two systems, while with MPIWasm, we observe a maximum bandwidth of 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='50 GiB/s and 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='83 GiB/s.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the collective communication Broadcast routine, we observe an average GM slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='13x with MPIWasm across all message sizes for 128 nodes as shown in Figure 3c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We observe an average GM slow- down of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='06x and 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='10x with MPIWasm across all message sizes for the collective communication AllReduce routine as shown in Figures 3d and 4c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For AllGather with MPI- Wasm, we observe an average GM slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='06x and 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='09x across all message sizes for the HPC system and Gravi- ton2 processor respectively (Figure 3e, 4d).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Similarly, for the Alltoall collective communication routine, we observe an average GM slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='10x for the two systems across all message sizes with MPIWasm as shown in Figures 3f and 4e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For 16 nodes of our HPC system, we observe an average GM slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='12x, 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='14x, and 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='05x across message sizes for the routines Reduce, Gather, and Scatter as shown in Figures 3g, 3h, and 3i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the other hand, for 128 nodes, PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 64128 256 512 1,024 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='8 1 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 104 Ranks Mop/s IS Total Operations/s Native WASM bh wh sh 0 500 1,000 1,500 Topology MB/s DT Total Throughput Native WASM w/o SIMD WASM w SIMD (a) NPB.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 4 8 12 16 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='8 3 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 104 Block Size (MiB) MiB/s IOR Bandwidth (Read) Native WASM 1 4 8 12 16 3 4 104 Block Size (MiB) MiB/s IOR Bandwidth (Write) Native WASM (b) IOR.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 48 16 48 96 144 0 20 40 60 80 Ranks GFLOP/s HPCG GFLOPS Native WASM 48 16 48 96 144 0 200 400 600 Ranks GB/s HPCG Bandwidth Native WASM 192 768 1,536 3,072 6,144 0 2,000 4,000 Ranks GFLOP/s HPCG GFLOPS Native WASM 192 768 1,536 3,072 6,144 0 1 2 3 104 Ranks GB/s HPCG Bandwidth Native WASM (c) HPCG.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Performance comparison of standardized HPC benchmarks for MPIWasm against their native execution on our HPC system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' we observe an average GM slowdown of 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='05x, 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='10x, and 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='08x for the three routines.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The results for testing MPIWasm with the Intel MPI Benchmarks compiled to Wasm demon- strate that neither the mechanism for calling host functions in Wasmer [87] nor the translation layer implemented in MPI- Wasm induce significant overhead for MPI communication (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5,§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6,§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We expand on the translation overhead in MPIWasm in §4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Overall, our results indicate that MPI- Wasm delivers close to native performance for the different MPI routines on both x86_64 and aarch64 architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The performance of MPIWasm on our HPC system for the IS and DT benchmarks is shown in Figure 5a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the IS benchmark with MPIWasm, we observe 8260 average mega operations per second across all processes as compared to 8546 average mega operations per second for the native exe- cution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the DT benchmark with different topologies (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2), execution with MPIWasm leads to decreased throughput as compared to the native execution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The DT benchmark per- forms a significant number of pairwise comparison opera- tions which benefit greatly from vectorization with SIMD instructions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To demonstrate the effect of SIMD for the DT benchmark, we compile it to Wasm by disabling and enabling the generation of SIMD instructions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm version of the DT benchmark with SIMD leads to 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='36x better throughput as compared to the Wasm version without SIMD (Figure 5a).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The difference in performance as compared to the native ver- sion of the DT benchmark can be attributed to the support for only 128-bit SIMD instructions in the Wasm specifica- tion [52] as compared to 512-bit SIMD instructions present in modern Intel processors [38, 57, 76] (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Support for higher-width SIMD in Wasm is an important milestone in its road-map but out of scope for this work (§5).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 5b shows total aggregated read and write band- width available to all MPI processes for the IOR benchmark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Points in the graph represent the average bandwidth reported by the benchmark, while error bars in the graph represent the maximum and minimum bandwidth observed over all iterations of the benchmark with the same block size.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With four nodes of our HPC system, the upper bound for achiev- able bandwidth in our setup (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1) with IOR is 400 GBit/s (≈ 47684 MiB/s).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With MPIWasm, we observe similar read (29411 MiB/s) and write (40206 MiB/s) bandwidth averaged across all block sizes as compared to the native execution of the benchmark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Testing the filesystem I/O performance of the MPIWasm demonstrates that the userspace permission handling and virtual directory tree implemented by Wasmer to provide filesystem isolation (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4) has no significant im- pact on the achievable bandwidth when performing I/O with the POSIX filesystem API.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For the HPCG benchmark, we ob- serve similar performance when executed with MPIWasm as compared to it’s native execution on the HPC system and the Graviton2 processor up to 192 MPI processes (Figures 5c and 4f).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On increasing the number of processes, the native exe- cution of the HPCG benchmark outperforms the execution with MPIWasm as shown in Figure 5c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For 6144 MPI processes, we observe a 14% reduction in GFLOP/s on execution with MPI- Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This behavior can be attributed to the significantly fre- quent amount of communication required by the HPCG bench- mark [63].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' HPCG repeatedly uses the Allreduce routine to reduce a single variable of size double over all MPI processes to finalize vector-vector dot operations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With increasing num- ber of processes, the number of times the Allreduce routine is called also increases.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For instance, executing HPCG with 768 processes results in four times more calls to Allreduce as compared to the execution with 192 processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a re- sult, the repeated datatype translations in MPIWasm increase the cost for invoking the collective communication routine leading to performance degradation (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 Analyzing Datatype Translation Overhead To measure the datatype translation overhead in MPIWasm, we implement a custom PingPong application that sends/re- ceives messages of varying sizes between two processes and iterates over the different MPI datatypes, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', BYTE, CHAR, INT, FLOAT, DOUBLE, and LONG.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, we instru- ment the Send routine in MPIWasm to determine the latency for translating the different datatypes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Finally, we execute the application on our HPC system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 6 shows the trans- lation overhead for different datatypes and message sizes in MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We observe an average overhead of 85.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='44ns, Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada 8 64 256 1024 32768 262144 1048576 2097152 4194304 Bytes 0 50 100 150 200 250 300 Translation Time (ns) MPI_BYTE MPI_CHAR MPI_INT MPI_FLOAT MPI_DOUBLE MPI_LONG Figure 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Comparing the datatype translation overhead in MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 84.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='72ns, 99.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='78ns, 96.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='32ns, 103.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='35ns, and 104.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='79ns across all message sizes for the MPI datatypes BYTE, CHAR, INT, FLOAT, DOUBLE, and LONG respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We observe an increase in the translation overhead for message sizes greater than 256KB.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This can be attributed to an increased latency for acquiring read locks from the Env structure that maintains the global state for translations in MPIWasm (§3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 5 Into the Future: Wasm and HPC In this section, we highlight and discuss the different exten- sions proposed by the Wasm community to the current Wasm specification [52] that can be implemented in an embedder for HPC applications to enhance performance and portability.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Controlled Threading for Wasm modules.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Wasm Threads proposal [16] lays the foundation for utilizing Wasm for multithreaded algorithms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It enables Wasm modules to de- fine shared memories, informing the embedder that the mod- ule expects the memory to be accessed by multiple threads.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To enable safe multithreaded access to shared memory, the proposal also defines atomic Wasm instructions that can be used to implement locks and atomic data structures in the functions of the module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To enable HPC applications to make use of the functionality added by the threads proposal, an API that allows Wasm modules to create additional threads on its own needs to be added to the embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Implementing the POSIX threads [15] and OpenMP [41] APIs in the embed- der would enable compatibility with the threading code in existing HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm Extended SIMD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The current Wasm SIMD pro- posal [52] only specifies 128-bit SIMD instructions, while modern CPUs support higher-width-SIMD, for instance the AVX-512 instruction set extensions for x86_64, which speci- fies 512-bit SIMD instructions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, the Wasm Flex- ible Vector proposal [13] aims to provide support for SIMD instructions that are wider than 128-bit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, the Wasm relaxed SIMD instructions [21] aim to make it possible to uti- lize hardware SIMD instructions that are not well defined, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', they differ in rounding behavior from the Wasm specification.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Implementing these proposals in the embedder would allow compiled Wasm modules for HPC applications that contain vectorizable code to make better use of SIMD instructions available in modern CPU architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 20 22 24 26 28 210 2 4 6 Bytes Iteration Time (usec) PingPong (time) ≤ 1024 Bytes MPIWasm Faasm 212 214 216 218 220 222 100 101 102 103 Bytes Iteration Time (usec) PingPong (time) > 1024 Bytes MPIWasm Faasm Figure 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Comparing the performance of MPIWasm and Faasm [78].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm PGAS.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Partitioned Global Address Space (PGAS) is a programming model for parallel distributed memory ap- plications that introduces a memory address space that spans the local memory of multiple processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With a memory ad- dress from this global address space a process can read from and write to the memory of other processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Since Wasm already specifies the concept of defining and importing mem- ories, the embedder could be extended to provide non-local memory to the Wasm module.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To support this use-case, the Wasm Multi-Memory proposal [14] needs to be implemented, which allows a Wasm module to define or import more than one memory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Dynamic Linking of Wasm Modules.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' While there is ex- isting work on establishing an ABI for dynamic linking be- tween Wasm modules [12], it has not been standardized yet.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Supporting dynamic linking would significantly decrease the size of Wasm binaries for more complex applications as they would no longer need to statically link parts of wasi-libc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For HPC, it would enable commonly used libraries such as BLAS to be provided by MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Combining dynamic link- ing with efforts to provide repositories for Wasm modules such as WAPM [88] could enable automatic dependency man- agement for Wasm applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Compiling Fortran applications to Wasm Currently, the support for compiling Fortran-based applications to Wasm is very nascent with only one known attempt based on Dragon- egg [17].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, the implementation of the Memory64 proposal [20] should enable the usage of existing Fortran LLVM compilers such as F18 for easily compiling Fortran- based applications to Wasm [18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm and Accelerators The module execution hints pro- posal [23] highlights the changes required in the Wasm speci- fication to enable the support for executing Wasm modules on hardware accelerators such as GPUs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Implementing the proposal in the embedder would enable compatibility with existing GPU-based HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 6 Related Work Solutions for Packaging and distributing HPC applica- tions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Recently several HPC-focused tools such as Char- liecloud [72] and Singularity [58] have been introduced for distributing HPC applications through containerization.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In contrast, we utilize the universal binary instruction format Wasm to package and distribute HPC applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' while containerization requires building HPC application con- tainers for different platforms, HPC applications can be com- piled once to Wasm and executed on any platform using a supporting Wasm embedder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI and WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To the best of our knowledge, Faasm [78] is the only compute platform that enables the exe- cution of MPI applications compiled to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It is based on a gRPC-based distributed messaging library called Faabric and contains a Wasm runtime, a workload scheduler, and a distributed state store.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In order to run an application on Faasm, it needs to be compiled to Wasm and uploaded to the shared function storage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, the application can be invoked using events such as HTTP requests.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For support- ing MPI applications, Faasm implements only a subset of the MPI-1 specification on top of its messaging library and its own workload scheduler.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Moreover, it also does not sup- port user-defined communicators required by the Intel MPI benchmarks (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In contrast, we take an inverted approach, where MPIWasm builds on top of existing native MPI libraries and provides a way for Wasm modules to call functions from them efficiently.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Figure 7 shows the performance comparison between MPIWasm and Faasm for the PingPong benchmark (§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' With MPIWasm, we achieve a GM average speedup of 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='28x across all message sizes as compared to Faasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 7 Conclusion and Future Work In this paper, we took the first step towards bringing We- bAssembly to the HPC ecosystem and presented MPIWasm, a Wasm embedder that enables high performance execution of MPI applications compiled to Wasm across different proces- sor architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In the future, we plan to extend MPIWasm to support acceralators such as GPUs found on HPC sys- tems [23].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 8 Acknowledgement We thank our shepherd Milind Chabbi for his help in prepar- ing the final version of this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Furthermore, we thank the anonymous reviewers for their insightful comments and valuable feedback that significantly improved the quality of this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This work was supported by the funding of the German Federal Ministry of Education and Research (BMBF) in the scope of the Software Campus program.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' References [1] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' AWS Graviton 2 Processors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://aws.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='amazon.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ec2/ graviton/ [2] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Blake-3 Hash function.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/BLAKE3-team/ BLAKE3 [3] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Cranelift Compiler.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/bytecodealliance/ wasmtime/tree/main/cranelift [4] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Intel Omni-Path Fabric.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='intel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/content/www/ us/en/high-performance-computing-fabrics/omni-path-fabric- software-components.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html [5] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Intel Performance Scaled Messaging 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ cornelisnetworks/opa-psm2/blob/master/README [6] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI: A Message-Passing Interface Standard.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mpi- forum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/docs/mpi-3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1/mpi31-report.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='pdf [7] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI bindings for Rust.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/rsmpi/rsmpi [8] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPICH: High-Performance portable MPI.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mpich.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' org/ [9] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MVAPICH:MPI over InfiniBand, Omni-Path, Ethernet/iWARP, and RoCE.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' http://mvapich.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='cse.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='ohio-state.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='edu/ [10] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' OpenMPI: Open-Source High-Performance computing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='open-mpi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/ [11] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The Top500 list.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='top500.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/lists/top500/2022/06/ [12] [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WebAssembly Dynamic Linking.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ WebAssembly/tool-conventions/blob/master/DynamicLinking.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [13] 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Flexible Vectors Proposal for WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' com/WebAssembly/flexible-vectors/blob/master/proposals/flexible- vectors/Overview.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [14] 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Multi Memory Proposal for WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' com/WebAssembly/multi-memory/blob/master/proposals/multi- memory/Overview.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [15] 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' POSIX Threads.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://man7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/linux/man-pages/man7/ pthreads.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html [16] 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Threading proposal for WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ WebAssembly/threads/blob/master/proposals/threads/Overview.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [17] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Compiling Fortran applications to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ StarGate01/Full-Stack-Fortran [18] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Fortran and WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://chrz.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='de/2020/04/21/fortran- in-the-browser/ [19] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Lenovo DSS-G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://tinyurl.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ytmjjauy [20] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Memory64 Proposal for WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/WebAssembly/memory64/blob/master/proposals/ memory64/Overview.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [21] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Relaxed SIMD Proposal for WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' com/WebAssembly/relaxed-simd/blob/main/proposals/relaxed- simd/Overview.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [22] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WASI Libc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/WebAssembly/wasi-libc [23] 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm and GPUs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/WebAssembly/design/ issues/1448 [24] 2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ARM Neoverse-N1 Architecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='arm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ products/silicon-ip-cpu/neoverse/neoverse-n1 [25] 2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Comparing Cranelift-IR and LLVM-IR.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/ bytecodealliance/wasmtime/blob/main/cranelift/docs/compare- llvm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md [26] 2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The GNU C Library (glibc).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='gnu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/software/libc/ [27] 2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The libloading library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='rs/libloading/latest/ libloading/ [28] 2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WASI SDK.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/WebAssembly/wasi-sdk [29] 14 Tech Companies Embracing Container Technology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //learn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='g2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/container-technology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/29/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [30] Subil Abraham, Arnab K Paul, Redwan Ibne Seraj Khan, and Ali R Butt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the use of containers in high performance comput- ing environments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2020 IEEE 13th International Conference on Cloud Computing (CLOUD).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' IEEE, 284–293.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/ CLOUD49709.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='00048 [31] D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Bailey, E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Barszcz, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Barton, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Browning, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Carter, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Dagum, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Fatoohi, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' O.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Frederickson, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Lasinski, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Schreiber, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Simon, V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Venkatakrishnan, and S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Weeratunga.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1991.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The NAS Parallel Benchmarks&Mdash;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='Summary and Prelim- inary Results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the 1991 ACM/IEEE Conference on Supercomputing (Albuquerque, New Mexico, USA) (Supercomputing ’91).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ACM, New York, NY, USA, 158–165.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/ 125826.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='125925 [32] JF Bastien, Thomas Lively, and Heejin Ahn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' LLVM WebAssembly Backend.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/llvm/llvm-project/blob/main/llvm/lib/ Target/WebAssembly/README.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='txt Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada [33] Lucas Benedicic, Felipe A Cruz, Alberto Madonna, and Kean Mariotti.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Sarus: Highly scalable docker containers for hpc systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In International Conference on High Performance Computing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Springer, 46–60.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1007/978-3-030-34356-9_5 [34] David E Bernholdt, Swen Boehm, George Bosilca, Manjunath Gorentla Venkata, Ryan E Grant, Thomas Naughton, Howard P Pritchard, Martin Schulz, and Geoffroy R Vallee.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A sur- vey of MPI usage in the US exascale computing project.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Concur- rency and Computation: Practice and Experience (2017).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1002/cpe.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4851 [35] Blazor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://dotnet.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='microsoft.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/apps/aspnet/web-apps/ blazor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [36] Brian Austin et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' NERSC-10 workload analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://portal.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='nersc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='gov/project/m888/nersc10/workload/N10_ Workload_Analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='latest.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='pdf.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [37] Bytecode Alliance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasmtime - A standalone runtime for We- bAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/bytecodealliance/wasmtime [38] Mohak Chadha, Anshul Jindal, and Michael Gerndt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Architecture- Specific Performance Optimization of Compute-Intensive FaaS Func- tions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2021 IEEE 14th International Conference on Cloud Computing (CLOUD).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 478–483.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/CLOUD53861.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='00062 [39] CMake.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://cmake.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed on 08/12/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [40] Compiling Rust to WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://developer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mozilla.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/ en-US/docs/WebAssembly/Rust_to_wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [41] Leonardo Dagum and Rameshm Enon.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1998.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' OpenMP: an industry standard API for shared-memory programming.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Computational Science & Engineering, IEEE 5, 1 (1998), 46–55.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [42] Frank Denis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Memory management in WebAssembly: guide for C and Rust programmers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='fastly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/blog/webassembly- memory-management-guide-for-c-rust-programmers [43] Docker.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='docker.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 07/07/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [44] Docker Buildx.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='docker.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/buildx/working-with- buildx/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 07/07/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [45] Emscripten.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://emscripten.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [46] FileSystemCache in Wasmer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='rs/wasmer-cache/3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2/wasmer_cache/struct.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='FileSystemCache.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 12/11/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [47] Phani Kishore Gadepalli, Sean McBride, Gregor Peach, Ludmila Cherkasova, and Gabriel Parmer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Sledge: A Serverless-First, Light-Weight Wasm Runtime for the Edge.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the 21st International Middleware Conference (Delft, Netherlands) (Middle- ware ’20).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, 265–279.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3423211.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3425680 [48] Holger Gantikow, Steffen Walter, and Christoph Reich.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Rootless Containers with Podman for HPC.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In International Conference on High Performance Computing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Springer, 343–354.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1007/ 978-3-030-59851-8_23 [49] Lisa Gerhardt, Wahid Bhimji, Shane Canon, Markus Fasel, Doug Jacobsen, Mustafa Mustafa, Jeff Porter, and Vakho Tsulaia.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Shifter: Containers for hpc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Journal of physics: Conference se- ries, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 898.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' IOP Publishing, 082021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1088/1742- 6596/898/8/082021 [50] GNU Make.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='gnu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/software/make/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed on 08/12/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [51] Robbert Gurdeep Singh and Christophe Scholliers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WARDuino: A Dynamic WebAssembly Virtual Machine for Programming Microcon- trollers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the 16th ACM SIGPLAN International Con- ference on Managed Programming Languages and Runtimes (Athens, Greece) (MPLR 2019).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, 27–36.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3357390.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3361029 [52] Andreas Haas, Andreas Rossberg, Derek L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Schuff, Ben L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Titzer, Michael Holman, Dan Gohman, Luke Wagner, Alon Zakai, and JF Bastien.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Bringing the Web up to Speed with WebAssembly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' SIGPLAN Not.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 52, 6 (June 2017), 185–200.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/ 3140587.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3062363 [53] Adam Hall and Umakishore Ramachandran.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' An Execution Model for Serverless Functions at the Edge.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the Interna- tional Conference on Internet of Things Design and Implementation (Montreal, Quebec, Canada) (IoTDI ’19).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, 225–236.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/ 3302505.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3310084 [54] Torsten Hoefler and Roberto Belli.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Scientific benchmarking of parallel computing systems: twelve ways to tell the masses when report- ing performance results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the international conference for high performance computing, networking, storage and analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1–12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/2807591.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2807644 [55] Intel Corporation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2018.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Introducing Intel® MPI Bench- marks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://software.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='intel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/content/www/us/en/develop/ articles/intel-mpi-benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html [56] Abhinav Jangda, Bobby Powers, Emery D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Berger, and Arjun Guha.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Not So Fast: Analyzing the Performance of WebAssembly vs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Native Code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2019 USENIX Annual Technical Conference (USENIX ATC 19).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' USENIX Association, Renton, WA, 107–120.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' usenix.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/conference/atc19/presentation/jangda [57] Michael Kiener, Mohak Chadha, and Michael Gerndt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To- wards Demystifying Intra-Function Parallelism in Serverless Comput- ing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the Seventh International Workshop on Server- less Computing (WoSC7) 2021 (Virtual Event, Canada) (WoSC ’21).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, 42–49.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3493651.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3493672 [58] Gregory M Kurtzer, Vanessa Sochat, and Michael W Bauer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Singularity: Scientific containers for mobility of compute.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' PloS one 12, 5 (2017), e0177459.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1371/journal.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='pone.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='0177459 [59] C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Lattner and V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Adve.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2004.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' LLVM: a compilation framework for lifelong program analysis transformation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In International Symposium on Code Generation and Optimization, 2004.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' CGO 2004.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 75–86.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/CGO.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2004.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1281665 [60] Lawrence Livermore National Laboratory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' HPC IO Benchmark Repository.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/hpc/ior [61] Linux.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Fakeroot.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://wiki.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='debian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/FakeRoot.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [62] Linux Kernel Support for Miscellaneous Binary Formats.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https:// www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='kernel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/doc/html/latest/admin-guide/binfmt-misc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Ac- cessed 07/07/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [63] Vladimir Marjanovi´c, José Gracia, and Colin W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Glass.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Per- formance Modeling of the HPCG Benchmark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In High Performance Computing Systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Performance Modeling, Benchmarking, and Simu- lation, Stephen A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Jarvis, Steven A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wright, and Simon D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Hammond (Eds.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Springer International Publishing, Cham, 172–192.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [64] Mellanox.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Infiniband.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://wiki.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='archlinux.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/title/ InfiniBand [65] Message Passing Interface Forum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2009.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI: A Message-Passing Interface Standard Version 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mpi-forum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/docs/mpi- 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2/mpi22-report.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='pdf [66] Message Passing Interface Forum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPI: A Message-Passing Interface Standard Version 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mpi-forum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/docs/mpi- 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1/mpi31-report.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='pdf [67] Mozilla.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' asm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='js.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' http://asmjs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/ [68] Mozilla.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Understanding WebAssembly text for- mat.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://developer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='mozilla.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/en-US/docs/WebAssembly/ Understanding_the_text_format [69] Open Container Initiative (OCI).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://opencontainers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 07/07/2022.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [70] Podman.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Filesystem Considerations in Rootless Mode.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //tinyurl.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/2p8efntt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [71] Reid Priedhorsky, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Shane Canon, Timothy Randles, and Andrew J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Younge.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Minimizing Privilege for Building HPC Containers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (St.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Louis, Missouri) PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' (SC ’21).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, Article 32, 14 pages.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3458817.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3476187 [72] Reid Priedhorsky and Tim Randles.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Charliecloud: Unprivileged containers for user-defined software stacks in hpc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the international conference for high performance computing, networking, storage and analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1–10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3126908.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3126925 [73] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Report, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Heroux, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Dongarra, and P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Luszczek.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2013.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' HPCG Technical Specification.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [74] Oleksandr Rudyy, Marta Garcia-Gasulla, Filippo Mantovani, Alfonso Santiago, Raül Sirvent, and Mariano Vázquez.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Containers in hpc: A scalability and portability study in production biological simulations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2019 IEEE International Parallel and Distributed Processing Sym- posium (IPDPS).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' IEEE, 567–577.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/IPDPS.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 00066 [75] Amit Ruhela, Stephen Lien Harrell, Richard Todd Evans, Gregory J Zynda, John Fonner, Matt Vaughn, Tommy Minyard, and John Cazes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Characterizing Containerized HPC Applications Performance at Petascale on CPU and GPU Architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In International Conference on High Performance Computing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Springer, 411–430.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/ 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1007/978-3-030-78713-4_22 [76] Robert Schöne, Thomas Ilsche, Mario Bielert, Andreas Gocht, and Daniel Hackenberg.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Energy efficiency features of the Intel Skylake-SP processor and their impact on performance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2019 Inter- national Conference on High Performance Computing & Simulation (HPCS).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' IEEE, 399–406.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/HPCS48598.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 9188239 [77] Thomas Schwinge.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' GCC Mailing List: Re: More compatibility it’s possible?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://gcc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='gnu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/legacy-ml/gcc/2019-11/msg00095.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' html [78] Simon Shillaker and Peter Pietzuch.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Faasm: Lightweight Isolation for Efficient Stateful Serverless Computing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2020 USENIX Annual Technical Conference (USENIX ATC 20).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' USENIX Association, 419– 433.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='usenix.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/conference/atc20/presentation/shillaker [79] Volodymyr Shymanskyy and Steven Massey.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/wasm3/wasm3 [80] Singularity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Filesystem Considerations in Rootless Mode.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://apptainer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/admin-docs/master/installation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [81] Solomon Hykes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Wasm+WASI: An alternative to Linux Con- tainers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://twitter.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/solomonstre/status/1111004913222324225?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' lang=en.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Accessed 09/27/2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [82] Michal Sudwoj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Rust programming language in the high- performance computing environment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' thesis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' ETH Zurich.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https: //www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='research-collection.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='ethz.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='ch/handle/20.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='500.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='11850/474922 [83] The kernel development community.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Seccomp BPF (SECure COMPuting with filters).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='kernel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/doc/html/v4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='16/ userspace-api/seccomp_filter.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='html [84] Alfred Torrez, Timothy Randles, and Reid Priedhorsky.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' HPC Container Runtimes have Minimal or No Performance Impact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2019 IEEE/ACM International Workshop on Containers and New Orchestra- tion Paradigms for Isolated Environments in HPC (CANOPIE-HPC).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 37–42.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/CANOPIE-HPC49598.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='00010 [85] Robert Wahbe, Steven Lucco, Thomas E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Anderson, and Susan L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Graham.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1993.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Efficient Software-Based Fault Isolation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' SIGOPS Oper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Syst.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 27, 5 (dec 1993), 203–216.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/173668.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 168635 [86] Wasmer, Inc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Single Pass Compiler in Wasmer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='rs/ wasmer-compiler-singlepass/latest/wasmer_compiler_singlepass/ [87] Wasmer, Inc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Universal WebAssembly Runtime.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://wasmer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' io/ [88] Wasmer, Inc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' [n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' wapm is the WebAssembly Package Manager.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://wapm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='io [89] WebAssembly Community Group.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' WebAssembly System Inter- face.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/WebAssembly/WASI [90] Miguel G Xavier, Marcelo V Neves, Fabio D Rossi, Tiago C Ferreto, Timoteo Lange, and Cesar AF De Rose.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2013.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Performance evalua- tion of container-based virtualization for high performance computing environments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In 2013 21st Euromicro International Conference on Parallel, Distributed, and Network-Based Processing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' IEEE, 233–240.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1109/PDP.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2013.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='41 [91] Yutian Yan, Tengfei Tu, Lijian Zhao, Yuchen Zhou, and Weihang Wang.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Understanding the Performance of Webassembly Applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Proceedings of the 21st ACM Internet Measurement Conference (Virtual Event) (IMC ’21).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Association for Computing Machinery, New York, NY, USA, 533–549.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1145/3487552.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3487827 [92] Andy B Yoo, Morris A Jette, and Mark Grondona.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2003.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Slurm: Simple linux utility for resource management.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' In Workshop on Job Scheduling Strategies for Parallel Processing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Springer, 44–60.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A Artifact Appendix A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 Description MPIWasm is an embedder for MPI-based HPC applications based on Wasmer [87].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' It enables the high performance execu- tion of these applications compiled to WebAssembly (Wasm) and serves two purposes: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Delivering close to native application performance, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', when applications are executed directly on a host ma- chine without using Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Enabling the distribution of MPI-based HPC applica- tions as Wasm binaries.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Our artifact contains the source code for MPIWasm, toolchain for compiling C/C++ based MPI applications to Wasm, the Wasm binaries for the different standardized HPC bench- marks used in this paper, pre-built versions of our embedder for different operating systems, and scripts for parsing experi- ment data and generating plots.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The artifact is available at: https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5281/zenodo.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7468121 or https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='com/kky-fury/MPIWasm A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 Getting Started For testing our Wasm embedder for executing MPI appli- cations compiled to WebAssembly, we provide a pre-built docker image for the linux/amd64 platform with all the re- quired dependencies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 sudo docker run − i t kkyfury / ppoppae : v2 / bin / bash 2 #Executing the HPCG benchmark compiled to Wasm 3 mpirun −−allow −run −as − r o o t −np 4 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / t a r g e t / r e l e a s e / embedder \\ 4 examples / xhpcg .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' wasm 5 #Executing the IntelMPI benchmarks compiled to Wasm 6 mpirun −−allow −run −as − r o o t −np 4 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / t a r g e t / r e l e a s e / embedder \\ 7 examples / imb .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' wasm Listing 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Getting started with MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Towards this, a user can follow the steps described in Listing 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, MPIWasm should successfully execute the HPCG and IntelMPI benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide sample output for the two benchmarks in the provided artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The user can increase/decrease the number of processes (-np) for executing the benchmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, depending on the system where Exploring the Use of WebAssembly in HPC PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada the container is executing, the user might need to provide the oversubscribe flag to mpirun.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3 Running Experiments with MPIWasm This section describes how to run experiments with our em- bedder to obtain plots similar to the ones in this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 Running small-scale experiments.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' To run small-scale experiments inside the docker container, we provide an end- to-end script.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This script: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Executes the HPCG, IS, and IntelMPI benchmarks for their native execution and when they are executed using MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Parses the obtained results and generates the relevant plots.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 sudo docker run − i t kkyfury / ppoppae : v2 / bin / bash 2 cd run_experiments 3 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / runme .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' sh Listing 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Running small-scale experiments with MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For running the experiments, the user can follow the steps de- scribed in Listing 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The script can take around 10-15 minutes to finish execution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' After completion, you can see the gener- ated data in the run_experiments/experiment_data folder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The generated plots can be found in the run_experiments/Plots folder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide sample plots for the different benchmarks in our artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' However, on executing benchmarks inside the container, the performance difference between the native execution of the application and MPIWasm can be around 8-12%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 Running large-scale experiments on an HPC sys- tem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For running large-scale experiments with our embedder, a user needs to do the following: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Build a version of the embedder for your HPC sys- tem depending on the particular architecture, operating system, and the MPI library on the system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MPIWasm currently supports the OpenMPI library with limited sup- port for MPICH and MVAPICH.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide examples for building MPIWasm for different operating systems and architectures in our artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Execute the MPI applications using the built embedder on the HPC system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This can be done via submitting jobs to a RJMS software on an HPC system such as SLURM [92].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide sample job scripts for our HPC system, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', SuperMUC-NG that uses SLURM in our artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' After executing the applications, the user can utilize the different parsers provided in our artifact to parse the benchmark data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Following this, the results can be visualized using the plotting helper provided in the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='4 Compiling C/C++ applications to Wasm We have setup a docker container with the required depen- dencies for compiling different MPI applications conformant to the MPI-2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 standard to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The artifact also includes HPCG, IntelMPI, and IS benchmarks as examples.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 sudo docker run − i t kkyfury / w a s i t o o l c h a i n : v1 / bin / bash 2 #Compiling HPCG 3 cd / work / example / hpcg −benchmark 4 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / wasi −cmake .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' sh 5 cd cmake− build −wasi 6 make Listing 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Compiling applications to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 7 describes the steps a user can follow for compiling the HPCG benchmark to Wasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For steps to compile the other benchmarks to Wasm, please look at the base Readme.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md file provided with the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' All the different applications compiled to Wasm that we used in this paper are also present in the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='5 Using MPIWasm For detailed usage instructions, please look at the base Readme.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md file provided with the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6 Modifying MPIWasm For modifying our embedder, we recommend using our pro- vided docker-compose file in the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' This docker-compose file mounts the volume with the embedder’s source code in- side the container.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' As a result, any changes to it’s source code will be reflected inside it.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' For our embedder, we currently support the following operating systems: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' CentOS-8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Opensuse-15-1 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Ubuntu-20-04 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' MacOS-monterey 1 cd wasi −mpi− r s 2 docker compose run centos −8−2 3 cargo b u i l d −− r e l e a s e 4 #After the build process , you can see the built embedder in 5 # the /target/release/ folder.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Building MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 8 describes the steps for building MPIWasm for CentOS-8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='2 after modifications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The instructions for building the embedder for other operating systems are provided in the base Readme.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='md file inside the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' After the build process, the embedder can be copied to the user’s local filesystem using the docker-cp command as shown in Listing 9.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 docker cp < c o n t a i n e r −id > : / s / t a r g e t / r e l e a s e / embedder \\ 2 < d e s t i n a t i o n −path −user − f i l e s y s t e m > Listing 9.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Copying MPIWasm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide provide the base image dockerfiles for the dif- ferent supported operating systems inside the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' These example dockerfiles can be easily extended to support other different linux distributions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' PPoPP ’23, February 25-March 1, 2023, Montreal, QC, Canada Mohak Chadha et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7 Support for aarch64 Our embedder also supports execution on linux/arm64 plat- forms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' We provide pre-built versions of our embedder for arm64 for the different supported operating systems in the artifact.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='1 Building images for aarch64.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' If the user is building the docker image on an x86_64 system then docker-buildx is required.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Note that, in this case, building the image might take around 12 hours.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 1 sudo docker buildx c r e a t e −−name mybuilder −−use −− b o o t s t r a p 2 cd wasi −mpi− r s / .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' g i t l a b / c i / images / 3 sudo docker buildx b u i l d −−push −f ubuntu −20 −04.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' D o c k e r f i l e \\ 4 −− p l at f or m l i n u x / arm64 \\ 5 − t kkyfury / ubuntumodifiedbase : v1 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' 6 cd .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' / 7 sudo docker buildx b u i l d −−push −f D o c k e r f i l e \\ 8 −− p l at f or m l i n u x / arm64 \\ 9 − t kkyfury / embedderarm : v1 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Building MPIWasm for aarch64 on x86_64.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' Listing 10 describes the steps required for building MPI- Wasm for arm systems with docker-buildx.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' The user should change the docker image tags according to their docker reg- istry account, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=', replace kkyfury with your registry user- name.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content=' On the other hand, if the user is using an aardch64 system then please follow the instructions described in §A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'} +page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/69E2T4oBgHgl3EQfkweU/content/2301.03982v1.pdf'}