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 14, 05:30 UTC
[XPU] Fix compile size for xpu (#23069)
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
Kunshang Ji committed
Aug 18, 2025 at 15:04 UTC
5f5664b3e4ff8046e26c36165a1294205cb429c5
1 file changed
+1
-1
vllm/config/__init__.py
+1
-1
@@ -3548,7 +3548,7 @@ class VllmConfig:
3548
if self.compilation_config.pass_config.enable_sequence_parallelism:
3549
self.compilation_config.custom_ops.append("+rms_norm")
3550
3551
- if current_platform.is_cuda_alike():
3551
+ if current_platform.is_cuda_alike() or current_platform.is_xpu():
3552
# if cudagraph_mode is not explicitly set by users, set default
3553
# value
3554
if self.compilation_config.cudagraph_mode is None: