Intel GPU D-Series Turns Crescent Island Into A 1.3 PFLOPS AI Beast
Intel has officially revealed the final form of Crescent Island, which turns out to be branded the "Intel GPU D-Series." The company's latest attempt to carve out a place in the data center AI accelerator market is led by the Intel GPU D270. This is a powerful inference-focused accelerator built around the new Xe3P architecture, supposedly the 'full' version of the Xe3 architecture that debuted with Panther Lake. The card combines up to 480GB of LPDDR5X memory with 1.3 PFLOPS of FP8 compute in a 350W power envelope, all in a conventional PCIe form factor with air cooling to simplify deployment in existing servers.
The big selling point for the D-Series is memory capacity, not speed, although it's no slouch. Intel says its GPU can accommodate models with up to 300 billion parameters on a single card, while an eight-GPU system can apparently handle models exceeding one trillion parameters....
Copyright of this story solely belongs to hothardware.com. To see the full text click HERE
