GAINESVILLE, Va., March 10, 2022 (GLOBE NEWSWIRE) — The Standard Performance Evaluation Corp. (SPEC) in the present day introduced the formation of the SPEC Machine Learning Committee. The SPEC ML Committee will develop sensible methodologies for benchmarking synthetic intelligence (AI) and machine studying (ML) efficiency within the context of real-world platforms and environments. The Committee may even work with different SPEC committees to replace their benchmarks for ML environments. Current members embrace AMD, Dell, Inspur, Intel, NetApp, NVIDIA and Red Hat. “IDC expects enterprises to spend almost $342 billion on AI in 2021, and it’s important that these corporations perceive what that cash will purchase,” stated Arthur Kang, Chair of the SPEC ML Committee. “This new committee will design and develop the seller agnostic benchmarks that distributors want to show their options in a aggressive market and enterprises want to make an knowledgeable shopping for determination. I encourage anybody fascinated with the way forward for ML processing to be a part of the SPEC ML Committee and assist form these invaluable benchmarks.” The SPEC ML Committee is initially creating benchmarks to measure end-to-end efficiency of a system underneath check (SUT) dealing with ML coaching and inference duties. The objective of those benchmarks is to higher characterize trade practices in contrast to different present benchmarks by together with main elements of the end-to-end ML/DL pipeline, together with knowledge prep and coaching/inference. This vendor-neutral third-party benchmark will allow ML system designers to benchmark their choices in opposition to these of their rivals and permit ML customers, reminiscent of enterprises and scientific analysis establishments, to higher perceive how options will carry out in real-world environments, enabling them to make higher buying selections. The SPEC ML Committee welcomes any present SPEC members prepared to be a part of the event and administration of the SPEC ML benchmark, in addition to new SPEC members, particularly ML/DL finish customers and producers. Visit membership or contact [email protected] for extra data. About SPEC SPEC is a non-profit group that establishes, maintains and endorses standardized benchmarks and instruments to consider efficiency for the latest era of computing techniques. Its membership contains greater than 120 main pc {hardware} and software program distributors, instructional establishments, analysis organizations, and authorities companies worldwide. Media contact: Brigit [email protected] Images out there upon request. SPEC® is a trademark of the Standard Performance Evaluation Corporation. All different product and firm names herein could also be logos of their registered house owners.
https://www.globenewswire.com/news-release/2022/03/10/2400558/0/en/SPEC-Establishes-Machine-Learning-Committee-to-Develop-Vendor-Agnostic-Benchmarks.html
