A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 06:00 UTC
[ROCm][V1] Adding ROCm to the list of plaforms using V1 by default (#19440)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
Gregory Shtrasberg committed
Jun 10, 2025 at 18:06 UTC
5241ca50d671e682c9e7fa2b70b761b0a7a4bc98
1 file changed
+1
-1
vllm/engine/arg_utils.py
+1
-1
@@ -1442,7 +1442,7 @@ class EngineArgs:
1442
1443
# Non-[CUDA, TPU] may be supported on V1, but off by default for now.
1444
v0_hardware = not any(
1445
- (current_platform.is_cuda(), current_platform.is_tpu(),
1445
+ (current_platform.is_cuda_alike(), current_platform.is_tpu(),
1446
(current_platform.is_cpu()
1447
and current_platform.get_cpu_architecture() == CpuArchEnum.X86)))
1448
if v0_hardware and _warn_or_fallback( # noqa: SIM103