Skip to content

GitLab

Unverified Commit 14d03b8d authored by 7. Sun's avatar 7. Sun Committed by GitHub
Browse files

[Perf] Cache xpu_get_mem_info() result to avoid duplicate calls (#32983) 


Signed-off-by: default avatar7. Sun <jhao.sun@gmail.com>
parent d0cbac58
Show whitespace changes
Inline Side-by-side
Showing with 2 additions and 1 deletion
vllm/v1/worker/xpu_worker.py
View file @ 14d03b8d
...@@ -107,7 +107,8 @@ class XPUWorker(Worker): ...@@ -107,7 +107,8 @@ class XPUWorker(Worker):
torch.xpu.empty_cache() torch.xpu.empty_cache()
torch_allocated_bytes = torch.xpu.memory_stats()["allocated_bytes.all.current"] torch_allocated_bytes = torch.xpu.memory_stats()["allocated_bytes.all.current"]
total_allocated_bytes = self.xpu_get_mem_info()[1] - self.xpu_get_mem_info()[0] free_mem, total_mem = self.xpu_get_mem_info()
total_allocated_bytes = total_mem - free_mem
non_torch_allocations = total_allocated_bytes - torch_allocated_bytes non_torch_allocations = total_allocated_bytes - torch_allocated_bytes
if non_torch_allocations > 0: if non_torch_allocations > 0:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment