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

[Hardware] Replace torch.cuda.device_count/current_device/set_device API (#36145)

Browse Source 
Signed-off-by: Kunshang Ji <jikunshang95@gmail.com>
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
This commit is contained in:
Kunshang Ji
2026-03-12 22:57:47 +08:00
committed by GitHub
parent 2e693f48e7
commit 53ec16a705
89 changed files with 254 additions and 219 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
benchmarks/kernels/benchmark_w8a8_block_fp8.py
View File

@@ -285,7 +285,7 @@ def tune_on_gpu(args_dict):
weight_shapes = args_dict["weight_shapes"]
args = args_dict["args"]
torch.cuda.set_device(gpu_id)
torch.accelerator.set_device_index(gpu_id)
print(f"Starting tuning on GPU {gpu_id} with batch sizes {batch_sizes}")
block_n = args.block_n
@@ -334,7 +334,7 @@ def distribute_batch_sizes(batch_sizes, num_gpus):
def main(args):
print(args)
num_gpus = torch.cuda.device_count()
num_gpus = torch.accelerator.device_count()
if num_gpus == 0:
raise RuntimeError("No GPU available for tuning")
print(f"Found {num_gpus} GPUs for parallel tuning")