CUDA compilation "No available targets are compatible with this triple." problem

Project Infrastructure LLVM Dev List Archives
1

Hi,

I have trouble compiling CUDA code with Clang. The following is a command I tried:

clang++ axpy.cu -o axpy --cuda-gpu-arch=sm_35 --cuda-path=/usr/local/cuda

The error message is

error: unable to create target: ‘No available targets are compatible with this triple.’

The info of the LLVM I’m using is as follows:

lang++ --version

clang version 6.0.0 (http://llvm.org/git/clang.git 16a0981eccf1bfcc9ba928747aa825e8948791cd) (http://llvm.org/git/llvm.git 052dd78cb30f77a05dc8bb06b851402c4b6c6587)

Target: x86_64-unknown-linux-gnu

Thread model: posix

InstalledDir: /home/xshen5/llvm/bin

Registered Targets:

x86 - 32-bit X86: Pentium-Pro and above

x86-64 - 64-bit X86: EM64T and AMD64

Any help I could get? Thanks!

-Xin

2

I’m assuming you are following guide at https://llvm.org/docs/CompileCudaWithLLVM.html#compiling-cuda-code

Have you tried -L and -l flags?

Kevin

3

Yes, I followed the guide. The same error showed up:

clang++ axpy.cu -o axpy --cuda-gpu-arch=sm_35 -L/usr/local/cuda/lib64 -I/usr/local/cuda/include -lcudart_static -ldl -lrt -pthread

error: unable to create target: ‘No available targets are compatible with this triple.’

4

If I am not mistaken, clang will compile CUDA code to PTX (which is then fed to ptxas), so I would assume you need to have the NVPTX target enabled, which does not seem to be the case for you.

Pierre

5

That was indeed the reason for the issue. The document of LLVM says that the default setting is all targets, which seems not true. I rebuilt LLVM with

cmake ~/clang-llvm/llvm -DCMAKE_BUILD_TYPE=Release -G Ninja -DLLVM_TARGETS_TO_BUILD=“X86;NVPTX”

The problem was then gone.
Thanks!