[Bug]: ROCm fail to build due to compilation error of moe_wna16.cu #14669
Description
Your current environment
The output of `python collect_env.py`
INFO 03-12 09:10:06 [__init__.py:256] Automatically detected platform rocm.
Collecting environment information...
PyTorch version: 2.7.0a0+git6c0e746
Is debug build: False
CUDA used to build PyTorch: N/A
ROCM used to build PyTorch: 6.3.42133-1b9c17779
OS: Ubuntu 22.04.5 LTS (x86_64)
GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Clang version: 18.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-6.3.1 24491 1e0fda770a2079fbd71e4b70974d74f62fd3af10)
CMake version: version 3.31.6
Libc version: glibc-2.35
Python version: 3.12.9 (main, Feb 5 2025, 08:49:00) [GCC 11.4.0] (64-bit runtime)
Python platform: Linux-5.15.0-116-generic-x86_64-with-glibc2.35
GPU models and configuration: AMD Instinct MI300X (gfx942:sramecc+:xnack-)
Nvidia driver version: Could not collect
cuDNN version: Could not collect
HIP runtime version: 6.3.42133
MIOpen runtime version: 3.3.0
Is XNNPACK available: True
PYTORCH_ROCM_ARCH=gfx942
LD_LIBRARY_PATH=/usr/local/lib/python3.12/dist-packages/cv2/../../lib64:/opt/rocm/lib:/usr/local/lib:
NCCL_CUMEM_ENABLE=0
TORCHINDUCTOR_COMPILE_THREADS=1
CUDA_MODULE_LOADING=LAZY
🐛 Describe the bug
When compiling the latest vLLM commit on ROCm, it gives the following error.
The error comes from the compilation of CUDA only kernels, introduced in 90e88ab.
Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.