cd/entity/vLLM-Gaudi· home› entities› vLLM-Gaudi
grep -l @vllm-gaudi /news/*.json | wc -l → 1

vLLM-Gaudi

mentions 1 type Organization feed RSS

// recent coverage 1 mentions

00:00
2026-09-01
sailresearch.com
artificial-intelligence

Improving Decode Throughput on Intel Gaudi 3

Intel Gaudi 3's vLLM-Gaudi PagedAttention implementation underutilizes hardware for sliding window attention, but optimizations for serving Gemma 4 31B expanded usable KV cache by 3.5×, raised decode …

// co-occurs with top 4 entities
Intel Gaudi 3 1 Gemma 4 31B 1 PagedAttention 1 Contiguous PagedAttention 1
// topics top 3 topics
artificial intelligence 1 ai infrastructure 1 ai research 1