AMD to introduce Zen 6-based EPYC CPUs at its Advancing AI Event beginning July 22

AMD is set to unveil its Zen 6 'Venice' EPYC server CPU lineup at its Advancing AI event in San Francisco on July 22 and 23.

According to an interview with The CUBE, AMD Chief Technology Officer and Executive Vice President, Mark Papermaster, confirmed that 6th Generation Zen CPUs will be introduced at the event. Venice is designed to enhance traditional enterprise x86 workloads, while also forming the backbone of Helios AI racks, paired with AMD MI455X GPUs.

Zen 6 EPYC processors have been in full production for some time. They're built with TSMC's 2nm process. It's believed to be the first high-performance product to enter production. Venice can scale up to 256 cores, with up to 1.6 TB/s of per-socket bandwidth. It brings PCIe 6.0 and 16-channel DDR5 support. It also introduces new AVX and VNNI extensions, making them well-suited to AI workloads.

3

Helios is AMD's rack-scale AI compute platform designed for cutting-edge LLM training and inference. It's designed to compete with NVIDIA's NVL72 racks. AMD touts its support for open standards, x86 architecture and memory capacity advantages over NVL72. But, the NVIDIA CUDA ecosystem remains dominant, and AMD has its work cut out if it wants to outperform the NVIDIA goliath.

As expected, AMD is allocating its production capacity for high-margin EPYC processors. Consumer Zen 6 is at least several months away, with a CES 2027 launch possible, though nothing is confirmed.

Frequently Asked Questions

TweakBot answers common questions about this news using TweakTown's own coverage from this page and related content from our archive. Tap a question to reveal the answer, or type your own below.

Question #1

How will the new AVX and VNNI extensions in Zen 6 EPYC affect performance for typical server AI inference workloads?

The primary article states Zen 6 EPYC introduces new AVX and VNNI extensions, which makes the processors "well-suited to AI workloads" and positions Venice as the backbone of Helios AI racks paired with MI455X GPUs. It does not provide quantitative performance details for typical server AI inference workloads, so the exact performance impact cannot be determined from this article.

Answered

Have a question not listed here? Ask below and TweakBot will answer it.

Your questionAsk

Answering your question×

TweakBot is analyzing your question…

Could not get an answer

TweakBot is generating an answer

Read More

Was TweakBot helpful?Thanks for the feedback!

The Advancing AI event kicks off on July 22, and we hope to get some more juicy info on Zen 6, even if consumers are going to be left hanging for some time to come.