Intel has provided new details on its new line of data center GPUs, dubbed Arctic Sound-M (ATS-M), first announced in February.
During Intel Vision 2022, the company revealed that there will be two ATS-M variants; a 150W card for maximum performance and a 75W card for maximum density.
Based on Intel's Xe-HPG microarchitecture (the same one that underpins the company's Arc chips), the new server GPUs feature 4 Xe Media Engines, up to 32 Xe cores, and a built-in AI accelerator. The company remains tight-lipped about the rest of the specs.
Arctic Sound-M for Visual Cloud
Intel's new data center GPUs are designed specifically for the visual cloud, a set of workloads and use cases that seek to deliver high pixel counts for purposes ranging from cloud gaming to virtual desktop infrastructure and AI inference.
During the first day, the lineup was characterized by Raja Koduri, the mastermind behind Intel's GPU efforts, as "a multimedia supercomputer on a single chip."
According to Koduri, the Arctic Sound-M chips are capable of handling more than 30 Full HD video streams, 40 cloud gaming streams, and 62 virtual desktops, and achieve up to 150 trillion AI operations per second.
Presumably, the new data center boards will also work synergistically with other Intel silicon boards, providing performance benefits not available with mixed configurations thanks to Intel's Deep Link technology suite.
The company's new Arc client cards, for example, can dynamically transfer power between the GPU and the CPU, depending on the demands of the specific workload.
Intel also wanted to highlight the importance of the open software stack supporting Arctic Sound-M, based on the oneAPI standards, which should make it easier for developers to build the platform.
The new data center cards are expected to launch inside servers from fifteen vendors, including Dell, Supermicro and HPE, in the third quarter of 2022.