Intel has officially named its Crescent Island AI accelerator the GPU D-Series and revealed its performance specifications. The chip is designed for AI inference tasks, offering 1.3 PFLOPS of FP8 compute and 1.5 TB/s of memory bandwidth. It will be available in an air-cooled PCIe form factor with a 350W TDP and up to 480GB of memory, targeting a different market than high-end solutions like Nvidia's Vera Rubin NVL72. AI
IMPACT Intel's new GPU D-Series aims to capture a share of the AI inference market with accessible form factors and competitive performance.
RANK_REASON Product launch announcement with detailed specifications from a major tech company.
Read on Mastodon — sigmoid.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →