ExOS Performance Summit

ExOS Performance Summit
Name	ExOS Performance Summit
Genre	Conference
Frequency	Annual
First	2019
Organiser	ExOS Foundation
Location	Global (rotating)

ExOS Performance Summit The ExOS Performance Summit is an annual conference focused on high-performance computing, systems optimization, and cloud-native infrastructure that convenes researchers, engineers, and industry leaders from organizations such as Intel Corporation, NVIDIA, Google LLC, Microsoft Corporation, and Amazon Web Services. The summit emphasizes practical benchmarking, reproducible research, and cross-sector collaboration among institutions like Massachusetts Institute of Technology, Stanford University, Lawrence Berkeley National Laboratory, Oak Ridge National Laboratory, and Los Alamos National Laboratory. It serves as a nexus for developers, vendors, and policymakers from entities including IEEE, ACM, The Linux Foundation, OpenStack Foundation, and Kubernetes communities.

Overview

The summit addresses performance engineering for platforms ranging from x86 servers and ARM architecture systems to GPU accelerators and FPGA deployments, bringing together stakeholders from AMD, ARM Limited, IBM, Cisco Systems, and Dell Technologies. Sessions cover topics with relevance to projects like TensorFlow, PyTorch, Hadoop, Spark (software), and standards bodies such as POSIX, OpenMP, MPI, OpenCL, and Vulkan (API). Attendees include representatives from research programs at National Institutes of Health, Defense Advanced Research Projects Agency, European Research Council, Japan Science and Technology Agency, and companies involved in edge computing such as Databricks, Red Hat, VMware, Cisco Systems, and Huawei.

History

Founded in 2019 by a coalition of performance engineers from organizations including Intel Corporation, Google LLC, AMD, NVIDIA, and The Linux Foundation, the summit evolved from workshops held at conferences like SC (conference), USENIX, OOPSLA, HotOS, and SIGCOMM. Early editions featured collaborations with laboratories such as Los Alamos National Laboratory, Oak Ridge National Laboratory, and Argonne National Laboratory, and partnered with initiatives like TOP500, Green500, SPEC (computer benchmark), MLPerf, and SPEC CPU. Over successive years it expanded to include industry partners like Microsoft Corporation, Amazon Web Services, IBM, Oracle Corporation, and regional research networks including GÉANT, JANET (UK), and SURFnet.

Program and Topics

The program combines peer-reviewed papers, hands-on tutorials, and benchmarking challenges covering areas such as kernel tuning for Linux kernel, compiler optimizations for GCC, Clang (compiler frontend), and LLVM, runtime systems for Java (programming language), Go (programming language), and Rust (programming language), as well as container orchestration with Kubernetes, service meshes like Istio, observability using Prometheus, Grafana, and tracing with Jaeger (software). Workshops explore storage stacks involving Ceph, ZFS, NVMe, and network fabrics like InfiniBand, RoCE, Ethernet, and protocol work on TCP, RDMA, and QUIC. Special tracks focus on machine learning performance with toolchains from TensorFlow, PyTorch, JAX, ONNX, and benchmarking suites such as MLPerf and SPEC.

Keynote Speakers and Participants

Keynotes have included executives and researchers from Intel Corporation, NVIDIA, Google LLC, Microsoft Research, Amazon Web Services, Facebook (Meta Platforms), ARM Limited, AMD, IBM Research, and academic leaders from Massachusetts Institute of Technology, Stanford University, University of California, Berkeley, Carnegie Mellon University, and University of Cambridge. Panels have featured contributors to projects and standards like Linux Foundation, OpenStack Foundation, Kubernetes, Hyperledger, and participants from consortia such as MLCommons and RISC-V International.

Research and Industry Contributions

The summit has catalyzed reproducible performance studies published by teams affiliated with Lawrence Berkeley National Laboratory, Argonne National Laboratory, Los Alamos National Laboratory, Oak Ridge National Laboratory, and universities including Princeton University, Harvard University, University of Oxford, and ETH Zurich. Industry collaborations have produced optimization patches upstreamed to Linux kernel, compiler improvements in GCC and LLVM, and benchmarking artifacts adopted by SPEC, MLPerf, and Top500 stakeholders. Corporate research groups from Intel Labs, NVIDIA Research, Google Research, Microsoft Research, Facebook AI Research, and IBM Research have used the summit to announce performance-driven initiatives and partnerships with cloud providers such as Amazon Web Services, Google Cloud Platform, Microsoft Azure, and Alibaba Cloud.

Awards and Recognition

The summit confers awards recognizing contributions to reproducible benchmarking, open-source performance tooling, and impactful optimization work, with honorees drawn from Linux Foundation, Apache Software Foundation, Eclipse Foundation, RISC-V International, and academic awardees from organizations like ACM and IEEE. Prize categories have included best paper awards, best artifact evaluations endorsed by Artifact Evaluation Committee panels, and industry innovation prizes sponsored by Intel Corporation, NVIDIA, Google LLC, Microsoft Corporation, and Amazon Web Services.

Venue and Organisation

The ExOS Performance Summit rotates venues including major conference centers in cities such as San Francisco, Seattle, Berlin, Tokyo, Paris, and London, and partners with academic hosts like Massachusetts Institute of Technology, Stanford University, ETH Zurich, and University of Tokyo. Organizational partners include The Linux Foundation, ACM SIGARCH, IEEE Computer Society, Open Source Initiative, and regional bodies like RIKEN, CNRS, and Deutsches Zentrum für Luft- und Raumfahrt.

Category:Computer performance conferences