biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Hardware] Replace torch.cuda.synchronize() api with torch.accelerator.synchronize (#36085)

Browse Source 
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
This commit is contained in:
Kunshang Ji
2026-03-05 18:36:39 +08:00
committed by GitHub
parent 0bfa229bf1
commit 66a2209645
59 changed files with 158 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
benchmarks/kernels/benchmark_trtllm_decode_attention.py
View File

@@ -126,7 +126,7 @@ def benchmark_decode(
)
def time_fn(fn, warmup=10, trials=20):
torch.cuda.synchronize()
torch.accelerator.synchronize()
start = torch.Event(enable_timing=True)
end = torch.Event(enable_timing=True)
times = []
@@ -136,7 +136,7 @@ def benchmark_decode(
start.record()
fn()
end.record()
torch.cuda.synchronize()
torch.accelerator.synchronize()
times.append(start.elapsed_time(end)) # ms
return sum(times) / len(times), torch.std(torch.tensor(times))