# vLLM speeds DeepSeek V4.1 lookup

> vLLM cuts DeepSeek V4.1 engram lookup from 54ms to 5.1ms on eight MI355X GPUs.
- Title: vLLM speeds DeepSeek V4.1 lookup · RuntimeWire | Shortify
- Summary: vLLM cuts DeepSeek V4.1 engram lookup from 54ms to 5.1ms on eight MI355X GPUs. vLLM reduced DeepSeek V4.1 engram lookup time from 54ms to 5.1ms using eight…
- Keywords: vllm, deepseek v4.1, engram lookup, mi355x, gpu, RuntimeWire
- URL: https://shortify.news/RuntimeWire/vllm-speeds-deepseek-v41-lookup
- Source: RuntimeWire — https://runtimewire.com/article/vllm-deepseek-v41-engram-lookup-mi355x
- Author: RuntimeWire
- Published: 2026-11-16T03:12:08.281+00:00
- Read time: 1 min
- Topics: vllm, deepseek v4.1, engram lookup, mi355x, gpu, technology, ai
## vLLM change
vLLM cut DeepSeek V4.1 engram lookup from 54 ms to 5.1 ms on eight MI355X GPUs.
## Fangzhou Ai fix
Fangzhou Ai traced DeepSeek V4.1 prefill stall on MI355X to GPU row fetch order.
## vLLM solution
vLLM ordered GPU requests before host memory to cut lookup time.
## Key takeaway

vLLM reduced DeepSeek V4.1 engram lookup time from 54ms to 5.1ms using eight MI355X GPUs.

---

Summary: Shortify — https://shortify.news/RuntimeWire/vllm-speeds-deepseek-v41-lookup
Reporting: RuntimeWire — https://runtimewire.com/article/vllm-deepseek-v41-engram-lookup-mi355x

When you use this summary, link the Shortify URL for the summary and credit the reporting to its newsroom.