biondizzle/grace-gpu-containers
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
cuda-malloc-managed
grace-gpu-containers/vllm/vllm_managed_mem.py
biondizzle bcc872c2c3 Remove global allocator swap, use targeted KV cache managed allocation 
sitecustomize.py: No longer swaps CUDAPluggableAllocator globally.
Sets VLLM_KV_CACHE_USE_MANAGED_MEMORY=1 instead.
vllm_managed_mem.py: No global allocator swap, no torch.cuda patches.
2026-04-11 02:15:09 +00:00

10 KiB
Raw Permalink Blame History

View Raw
Reference in New Issue View Git Blame Copy Permalink