Intel Announces D-Series GPUs for AI Data Centers
Intel's new D270 GPU offers up to 1.3 PFLOPs of FP8 compute power.
Intel has announced its D-Series GPUs, previously known as Crescent Island, as the next lineup for AI data centers based on the Xe3P architecture. The first product in this series is the Intel GPU D270.
The D270 delivers up to 1.3 PFLOPs of FP8 compute at a TDP of 350W. It supports LPDDR5X memory with capacities ranging from 160GB (reference) to 480GB (partner models). The GPU contains 32 Xe3P cores, 256 XMX engines, and 32MB of unified L2 cache.
Key facts
- Intel D-Series GPUs announced for AI data centers
- D270 delivers up to 1.3 PFLOPs FP8 compute
- TDP of 350W
- Supports LPDDR5X memory (160GB-480GB)
- Contains 32 Xe3P cores and 256 XMX engines
Sources
- WccftechIntel Crescent Island Squeezes 1.3 PFLOPs Of FP8 Into Just 350W As D270 Targets The Competition With Value-Oriented LPDDR5X Up To 480 GB
- TechPowerUpIntel "Crescent Island" GPU Gets D-Series Name, Supports Up to 480 GB of LPDDR5X