Skip to content

perf(amd): Kimi-K3.1 MI355X vLLM agentic: 898 GB CPU offload, mbt 8192, capture cap 8192 - #4867

Closed
seungrokj wants to merge 33 commits into
mainfrom
srok/kimik3.1-mi355x-vllm-cpu2700-mbt8k
Closed

seungrokj wants to merge 33 commits into
mainfrom
srok/kimik3.1-mi355x-vllm-cpu2700-mbt8k

Commits

  1. Commits on Nov 4, 2026

  2. Commits on Nov 5, 2026

  3. Commits on Nov 6, 2026