Closed
Description
About libc for GPU, OpenMP offload example does not succeed.
I tried the following script to build clang.
cmake /volume/develop/src/llvm-project/llvm -G Ninja \
-DCMAKE_BUILD_TYPE=Release \
-DLLVM_ENABLE_PROJECTS="clang;lld" \
-DLLVM_ENABLE_RUNTIMES="openmp" \
-DCMAKE_INSTALL_PREFIX=/path to install \
-DCMAKE_C_COMPILER=/path to clang \
-DCMAKE_CXX_COMPILER=/path to clang++ \
-DRUNTIMES_nvptx64-nvidia-cuda_LLVM_ENABLE_RUNTIMES=libc \
-DRUNTIMES_amdgcn-amd-amdhsa_LLVM_ENABLE_RUNTIMES=libc \
-DLLVM_RUNTIME_TARGETS="default;amdgcn-amd-amdhsa;nvptx64-nvidia-cuda"
However, the following procedure resulted in an error.
https://libc.llvm.org/gpu/using.html#openmp-offloading-example
$ clang -v openmp.c -fopenmp --offload-arch=sm_86
clang version 19.1.7 (https://github.com/llvm/llvm-project.git cd708029e0b2869e80abe31ddb175f7c35361f90)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /--/install/bin
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/11
Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/11
Candidate multilib: .;@m64
Selected multilib: .;@m64
Found CUDA installation: /usr/local/cuda-12.1, version 12.1
clang: error: no library 'libomptarget-nvptx-sm_86.bc' found in the default clang lib directory or in LIBRARY_PATH; use '--libomptarget-nvptx-bc-path' to specify nvptx bitcode library
I couldn't find libomptarget-nvptx-sm_*.bc, but found libc.bc.
./lib/amdgcn-amd-amdhsa/libc.bc
./lib/amdgcn-amd-amdhsa/libm.bc
./lib/nvptx64-nvidia-cuda/libc.bc
./lib/nvptx64-nvidia-cuda/libm.bc
CPU Intel Xeon Gold 6338T
GPU NVIDIA RTX A6000
OS Ubuntu 22.04
LLVM version 19.1.7
NVidia Driver 545.23.06
CUDA 12.1