MegEngine / MegEngine/InferLLM
能否改为GPU辅助计算
- Dominant language
- C++
- Stars
- 752
- Forks
- 94
- PR merge metrics
- No merged PRs in 30d
Description
`bzy@server:~/InferLLM/build$ make
-- current platform: x86
-- Build with GPU.
-- Configuring done
-- Generating done
-- Build files have been written to: /home/bzy/InferLLM/build
[ 2%] Building NVCC (Device) object CMakeFiles/InferLLMGPU.dir/src/kern/gpu/InferLLMGPU_generated_kernel_gpu.cu.o
/usr/lib/gcc/x86_64-linux-gnu/11/include/serializeintrin.h(41): error: identifier "__builtin_ia32_serialize" is undefined
/home/bzy/InferLLM/src/kern/naive/quantize.h(45): warning #177-D: variable "bs" was declared but never referenced
1 error detected in the compilation of "/home/bzy/InferLLM/src/kern/gpu/kernel_gpu.cu".
CMake Error at InferLLMGPU_generated_kernel_gpu.cu.o.cmake:280 (message):
Error generating file
/home/bzy/InferLLM/build/CMakeFiles/InferLLMGPU.dir/src/kern/gpu/./InferLLMGPU_generated_kernel_gpu.cu.o
make[2]: *** [CMakeFiles/InferLLMGPU.dir/build.make:522:
CMakeFiles/InferLLMGPU.dir/src/kern/gpu/InferLLMGPU_generated_kernel_gpu.cu.o] 错误 1
make[1]: *** [CMakeFiles/Makefile2:94:CMakeFiles/InferLLMGPU.dir/all] 错误 2
make: *** [Makefile:91:all] 错误 2
`
Contributor guide
No contributing guide indexed for this repository
Research direction
Reproduce the GPU build from the `build` directory with `make`, then inspect `src/kern/gpu/kernel_gpu.cu` and the reported compiler failure involving `/usr/lib/gcc/x86_64-linux-gnu/11/include/serializeintrin.h`. Also review `src/kern/naive/quantize.h` for the warning. Done means the GPU target configures and compiles without the reported NVCC error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cmake, cpp
- Domain
- build-system, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100