AMD launches 5th gen EPYC CPUs to supercharge enterprise AI, cloud workloads

October 10, 2024 07:00 PM

AMD has released its fifth generation EPYC CPUs, dubbed Turin, to power intense computing workloads for enterprise AI and cloud workloads.

Just one fifth-generation EPYC server is powerful enough that it could replace up to seven legacy servers, Dr. Lisa Su, CEO of AMD suggested during the Advancing AI keynote.

Subscribe today for free

The connectivity news and insights that matter - straight to your inbox

Built on the new “Zen 5” core architecture, the new EPYC CPU servers provide up to 17% better instructions per clock (IPC) for enterprise and cloud workloads compared to the previous generation hardware.

Su said AMD’s EPYC line has become “the CPU of choice for the modern data centre.”

“The largest cloud providers offer more than 950 EPYC instances and have deployed epic widely throughout their infrastructure,” Su said. “And on the enterprise side, numerous large customers have also deployed EPYC on-prem to power their most important workloads.”

The new CPU comes in a broad range of sizes, ranging from core counts from eight up to 192.

AMD claims its top-end 192 core unit delivers up to 2.7 times the performance compared to the rival CPUs and up to 3.7 times the performance on performance on end-to-end AI workloads like TPCx-AI which are critical for driving an efficient approach to generative AI

The hardware giant suggested the EPYC 9965 processor-based servers offer up to four times faster time to results on business applications such as video transcoding, and Up to 3.9X the time to insights for science and HPC applications compared to Intel’s Xeon 8592+ CPU-based servers.

“With five generations of on-time roadmap execution, AMD has proven it can meet the needs of the data centre market and give customers the standard for data centre performance, efficiency, solutions and capabilities for cloud, enterprise and AI workloads,” said Dan McNamara, SVP and general manager of AMD’s server business.

Google Cloud is among the early adopters of the fifth-generation EPYC servers with plans to hardware to power virtual machine solutions in 2025.

“EPYC CPUs are an important part of that stack, offering a cost-effective and seamless option for AI workloads,” said Amin Vahdat, VP and general manager for machine learning, systems, and cloud AI at Google Cloud.