Unsupported gpu architecture 'compute_'

[swpper]

Hi, I am learning llama.cpp, and I met a simple problem when excute cmake -B build_cuda -DLLAMA_CUDA=ON, Below is the output.

versions:

cuda: 12.1
llama.cpp: latest master branch

vllm
unsloth
python: 3.10
llama-factory
...

(base) root@autodl-container-8fd941ac7c-d999e938:~/autodl-tmp/llama.cpp# cmake -B build_cuda -DLLAMA_CUDA=ON
CMake Warning at CMakeLists.txt:102 (message):
  LLAMA_CUDA is deprecated and will be removed in the future.

  Use GGML_CUDA instead

Call Stack (most recent call first):
  CMakeLists.txt:108 (llama_option_depr)


-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with GGML_CCACHE=OFF
-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- OpenMP found
-- Using llamafile
-- x86 detected
-- Using runtime weight conversion of Q4_0 to Q4_0_x_x to enable optimized GEMM/GEMV kernels
-- Including CPU backend
-- Using AMX
-- Including AMX backend
-- CUDA Toolkit found
-- Using CUDA architectures: native
CMake Error at /usr/share/cmake-3.22/Modules/CMakeDetermineCompilerId.cmake:726 (message):
  Compiling the CUDA compiler identification source file
  "CMakeCUDACompilerId.cu" failed.

  Compiler: /usr/local/cuda/bin/nvcc

  Build flags:

  Id flags: --keep;--keep-dir;tmp;-gencode=arch=compute_,code=sm_ -v

  

  The output was:

  1

  nvcc fatal : Unsupported gpu architecture 'compute_'

  

  

Call Stack (most recent call first):
  /usr/share/cmake-3.22/Modules/CMakeDetermineCompilerId.cmake:6 (CMAKE_DETERMINE_COMPILER_ID_BUILD)
  /usr/share/cmake-3.22/Modules/CMakeDetermineCompilerId.cmake:48 (__determine_compiler_id_test)
  /usr/share/cmake-3.22/Modules/CMakeDetermineCUDACompiler.cmake:298 (CMAKE_DETERMINE_COMPILER_ID)
  ggml/src/ggml-cuda/CMakeLists.txt:25 (enable_language)


-- Configuring incomplete, errors occurred!
See also "/root/autodl-tmp/llama.cpp/build_cuda/CMakeFiles/CMakeOutput.log".
See also "/root/autodl-tmp/llama.cpp/build_cuda/CMakeFiles/CMakeError.log".

I think it is a simpe problem but I don`t know what it is. Can you give me a help?

[slaren]

It should be fixed now.

[swpper]

    Hello, I have received your message in my email and will response quickly, thank you.

[manojneversu]

i have the same error, trying to fix, but cant be able to, please help me out

[swpper]

Hi, I am no longer deep in llama.cpp from that time, unfortunately I forgot how to solve it, sorry.

[swpper]

    Hello, I have received your message in my email and will response quickly, thank you.

[manojneversu]

Worked it around bruh! Thanks for the reply tho!

…

On Thu, Jun 19, 2025, 4:21 PM swpper ***@***.***> wrote: Hi, I am no longer deep in llama.cpp from that time, unfortunately I forgot how to solve it, sorry. — Reply to this email directly, view it on GitHub <#10388 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BNFN6GOGPB2N6SJJ4QC4JTL3EKI35AVCNFSM6AAAAABSACTHBKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNJRHEYDEOA> . You are receiving this because you commented.Message ID: <ggml-org/llama. ***@***.***>