Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Workaround to build with CUDA #644

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
ServeurpersoCom opened this issue Apr 2, 2025 · 2 comments
Open

Workaround to build with CUDA #644

ServeurpersoCom opened this issue Apr 2, 2025 · 2 comments

Comments

@ServeurpersoCom
Copy link

After a git submodule update (ggml) on main :

-- The C compiler identification is GNU 10.2.1
-- The CXX compiler identification is GNU 10.2.1
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Build static library
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Check if compiler accepts -pthread
-- Check if compiler accepts -pthread - yes
-- Found Threads: TRUE
-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with GGML_CCACHE=OFF
-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- Including CPU backend
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- x86 detected
-- Adding CPU backend variant ggml-cpu: -march=native
-- Found BLAS: /root/aocl/5.0.0/gcc/lib/libblis.so
-- BLAS found, Libraries: /root/aocl/5.0.0/gcc/lib/libblis.so
-- BLAS found, Includes: /root/aocl/5.0.0/gcc/include
-- Including BLAS backend
-- Found CUDAToolkit: /usr/local/cuda/include
-- CUDA Toolkit found
-- Using CUDA architectures: 50;61;70;75;80
-- The CUDA compiler identification is NVIDIA 11.8.89
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- CUDA host compiler is GNU 10.2.1

-- Including CUDA backend
-- Configuring done
-- Generating done
-- Build files have been written to: /var/www/stats/stable-diffusion.cpp/build
Scanning dependencies of target ggml-base
Scanning dependencies of target zip
[  1%] Building C object ggml/src/CMakeFiles/ggml-base.dir/ggml-quants.c.o
[  2%] Building C object ggml/src/CMakeFiles/ggml-base.dir/ggml.c.o
[  5%] Building C object ggml/src/CMakeFiles/ggml-base.dir/ggml-alloc.c.o
[  5%] Building CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-backend.cpp.o
[  5%] Building CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-opt.cpp.o
[  6%] Building CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-threading.cpp.o
[  7%] Building C object thirdparty/CMakeFiles/zip.dir/zip.c.o
[  7%] Building CXX object ggml/src/CMakeFiles/ggml-base.dir/gguf.cpp.o
In file included from /var/www/stats/stable-diffusion.cpp/thirdparty/zip.c:40:
/var/www/stats/stable-diffusion.cpp/thirdparty/miniz.h:4988:9: note: ‘#pragma message: Using fopen, ftello, fseeko, stat() etc. path for file I/O - this path may not support large files.’
 4988 | #pragma message(                                                               \
      |         ^~~~~~~
[  8%] Linking CXX static library libggml-base.a
[  8%] Built target ggml-base
Scanning dependencies of target ggml-cpu
Scanning dependencies of target ggml-blas
Scanning dependencies of target ggml-cuda
[ 10%] Building C object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu.c.o
[ 10%] Building CXX object ggml/src/ggml-blas/CMakeFiles/ggml-blas.dir/ggml-blas.cpp.o
[ 11%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/mmq.cpp.o
[ 12%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu.cpp.o
[ 13%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-aarch64.cpp.o
[ 15%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-traits.cpp.o
[ 15%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-hbm.cpp.o
[ 16%] Building C object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-quants.c.o
[ 16%] Building CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/amx.cpp.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/mmv.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/arange.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/cpy.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/acc.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/gla.cu.o
[ 28%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/fattn-wmma-f16.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/ggml-cuda.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/im2col.cu.o
[ 28%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/concat.cu.o
[ 28%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/getrows.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/fattn-tile-f32.cu.o
[ 30%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/binbcast.cu.o
[ 30%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/convert.cu.o
[ 32%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/argmax.cu.o
[ 32%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/conv-transpose-1d.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/count-equal.cu.o
[ 33%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/mmq.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/fattn.cu.o
[ 27%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/clamp.cu.o
[ 34%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/cross-entropy-loss.cu.o
[ 35%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/argsort.cu.o
[ 35%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/fattn-tile-f16.cu.o
[ 36%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/diagmask.cu.o
[ 37%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/mmvq.cu.o
[ 38%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/norm.cu.o
[ 39%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/opt-step-adamw.cu.o
[ 40%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/out-prod.cu.o
[ 41%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/pad.cu.o
[ 42%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/pool2d.cu.o
[ 43%] Linking CXX static library libggml-cpu.a
[ 44%] Linking CXX static library libggml-blas.a
[ 44%] Built target ggml-cpu
[ 45%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/quantize.cu.o
[ 45%] Built target ggml-blas
[ 46%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/rope.cu.o
/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu(11): error: more than one conversion function from "const half" to a built-in type applies:
            function "__half::operator float() const"
/usr/local/cuda/include/cuda_fp16.hpp(204): here
            function "__half::operator short() const"
/usr/local/cuda/include/cuda_fp16.hpp(222): here
            function "__half::operator unsigned short() const"
/usr/local/cuda/include/cuda_fp16.hpp(225): here
            function "__half::operator int() const"
/usr/local/cuda/include/cuda_fp16.hpp(228): here
            function "__half::operator unsigned int() const"
/usr/local/cuda/include/cuda_fp16.hpp(231): here
            function "__half::operator long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(234): here
            function "__half::operator unsigned long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(237): here
            function "__half::operator __nv_bool() const"
/usr/local/cuda/include/cuda_fp16.hpp(241): here
          detected during:
            instantiation of "void op_clamp(const T *, T *, T, T, int) [with T=half]"
(17): here
            instantiation of "void clamp_cuda(const T *, T *, T, T, int, cudaStream_t) [with T=half]"
(37): here

/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu(11): error: more than one conversion function from "const half" to a built-in type applies:
            function "__half::operator float() const"
/usr/local/cuda/include/cuda_fp16.hpp(204): here
            function "__half::operator short() const"
/usr/local/cuda/include/cuda_fp16.hpp(222): here
            function "__half::operator unsigned short() const"
/usr/local/cuda/include/cuda_fp16.hpp(225): here
            function "__half::operator int() const"
/usr/local/cuda/include/cuda_fp16.hpp(228): here
            function "__half::operator unsigned int() const"
/usr/local/cuda/include/cuda_fp16.hpp(231): here
            function "__half::operator long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(234): here
            function "__half::operator unsigned long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(237): here
            function "__half::operator __nv_bool() const"
/usr/local/cuda/include/cuda_fp16.hpp(241): here
          detected during:
            instantiation of "void op_clamp(const T *, T *, T, T, int) [with T=half]"
(17): here
            instantiation of "void clamp_cuda(const T *, T *, T, T, int, cudaStream_t) [with T=half]"
(37): here

[ 46%] Built target zip
/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu(11): error: more than one conversion function from "const half" to a built-in type applies:
            function "__half::operator float() const"
/usr/local/cuda/include/cuda_fp16.hpp(204): here
            function "__half::operator short() const"
/usr/local/cuda/include/cuda_fp16.hpp(222): here
            function "__half::operator unsigned short() const"
/usr/local/cuda/include/cuda_fp16.hpp(225): here
            function "__half::operator int() const"
/usr/local/cuda/include/cuda_fp16.hpp(228): here
            function "__half::operator unsigned int() const"
/usr/local/cuda/include/cuda_fp16.hpp(231): here
            function "__half::operator long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(234): here
            function "__half::operator unsigned long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(237): here
            function "__half::operator __nv_bool() const"
/usr/local/cuda/include/cuda_fp16.hpp(241): here
          detected during:
            instantiation of "void op_clamp(const T *, T *, T, T, int) [with T=half]"
(17): here
            instantiation of "void clamp_cuda(const T *, T *, T, T, int, cudaStream_t) [with T=half]"
(37): here

/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu(11): error: more than one conversion function from "const half" to a built-in type applies:
            function "__half::operator float() const"
/usr/local/cuda/include/cuda_fp16.hpp(204): here
            function "__half::operator short() const"
/usr/local/cuda/include/cuda_fp16.hpp(222): here
            function "__half::operator unsigned short() const"
/usr/local/cuda/include/cuda_fp16.hpp(225): here
            function "__half::operator int() const"
/usr/local/cuda/include/cuda_fp16.hpp(228): here
            function "__half::operator unsigned int() const"
/usr/local/cuda/include/cuda_fp16.hpp(231): here
            function "__half::operator long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(234): here
            function "__half::operator unsigned long long() const"
/usr/local/cuda/include/cuda_fp16.hpp(237): here
            function "__half::operator __nv_bool() const"
/usr/local/cuda/include/cuda_fp16.hpp(241): here
          detected during:
            instantiation of "void op_clamp(const T *, T *, T, T, int) [with T=half]"
(17): here
            instantiation of "void clamp_cuda(const T *, T *, T, T, int, cudaStream_t) [with T=half]"
(37): here

/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu(11): error: ambiguous "?" operation: second operand of type "const half" can be converted to third operand type "<error-type>", and vice versa
          detected during:
            instantiation of "void op_clamp(const T *, T *, T, T, int) [with T=half]"
(17): here
            instantiation of "void clamp_cuda(const T *, T *, T, T, int, cudaStream_t) [with T=half]"
(37): here

[ 46%] Building CUDA object ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/scale.cu.o
5 errors detected in the compilation of "/var/www/stats/stable-diffusion.cpp/ggml/src/ggml-cuda/clamp.cu".
gmake[2]: *** [ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/build.make:147 : ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/clamp.cu.o] Erreur 1
gmake[2]: *** Attente des tâches non terminées....
gmake[1]: *** [CMakeFiles/Makefile2:402 : ggml/src/ggml-cuda/CMakeFiles/ggml-cuda.dir/all] Erreur 2
gmake: *** [Makefile:149 : all] Erreur 2
@ServeurpersoCom
Copy link
Author

OK I get the last ggml submodule from llama.cpp and CUDA build it work on stable-diffusion.cpp :)
submodule on this project need update :)

@ServeurpersoCom
Copy link
Author

Now we net to checkout to 8b9cc7cdd8a0dcf0176c60c755322c95b5965299 to get the last working ggml because Georgi work on it

@ServeurpersoCom ServeurpersoCom changed the title Can't build with CUDA Workaround to build with CUDA Apr 12, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@ServeurpersoCom