kvcache-ai/ktransformers
1
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2026-11-05 01:07:28 +00:00
Code Issues 469 Projects Packages Wiki Activity

[GH-ISSUE #984] [Bug] 不支持海光DCU K100A （gfx928） #698

New issue
Closed
opened 2026-10-29 14:43:00 +00:00 by bsakaguchi · 4 comments
bsakaguchi commented 2026-10-29 14:43:00 +00:00
Owner
Copy link

Originally created by @yeungtuzi on GitHub (Mar 27, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/984

检查清单

  • 1. 我已经搜索过相关问题，但未能获得预期的帮助
  • 2. 该问题在最新版本中尚未修复
  • 3. 请注意，如果您提交的BUG相关 issue 缺少对应环境信息和最小可复现示例，我们将难以复现和定位问题，降低获得反馈的可能性
  • 4. 如果您提出的不是bug而是问题，请在讨论区发起讨论 https://github.com/kvcache-ai/ktransformers/discussions。否则该 issue 将被关闭
  • 5. 为方便社区交流，我将使用中文/英文或附上中文/英文翻译（如使用其他语言）。未附带翻译的非中文/英语内容可能会被关闭

问题描述

在make dev_install的时候报错

  [54/54] : && /usr/bin/c++ -fPIC -O3 -ffast-math -O3 -DNDEBUG  -flto -shared  -o /tmp/tmpuqdwl1qf.build-lib/cpuinfer_ext.cpython-311-x86_64-linux-gnu.so CMakeFiles/cpuinfer_ext.dir/ext_bindings.cpp.o CMakeFiles/cpuinfer_ext.dir/cpu_backend/backend.cpp.o CMakeFiles/cpuinfer_ext.dir/cpu_backend/task_queue.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/linear.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/mlp.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/moe.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/shared_mem_buffer.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/flags.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/sgemm.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx512f.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avxvnni.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_fma.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_arm80.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx512f.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avxvnni.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_fma.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_arm80.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_unsupported.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_attn.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_load_dump.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_read_write.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_utils.cpp.o  -Wl,-rpath,/usr/local/cuda/lib64  third_party/llama.cpp/libllama.a  /usr/local/cuda/lib64/libcudart.so  /usr/lib/gcc/x86_64-linux-gnu/11/libgomp.so  /usr/lib/x86_64-linux-gnu/libpthread.a && cd /root/ktransformers/ktransformers/ktransformers_ext/build && /usr/bin/strip /tmp/tmpuqdwl1qf.build-lib/cpuinfer_ext.cpython-311-x86_64-linux-gnu.so
  lto-wrapper: warning: using serial compilation of 15 LTRANS jobs
  CMake args: ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/tmp/tmpuqdwl1qf.build-lib/', '-DPYTHON_EXECUTABLE=/opt/anaconda3/envs/kt/bin/python', '-DCMAKE_BUILD_TYPE=Release', '-DKTRANSFORMERS_USE_CUDA=ON']
  CMake args: ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/tmp/tmpuqdwl1qf.build-lib/', '-DPYTHON_EXECUTABLE=/opt/anaconda3/envs/kt/bin/python', '-DCMAKE_BUILD_TYPE=Release', '-DKTRANSFORMERS_USE_CUDA=ON', '-DLLAMA_NATIVE=ON', '-DEXAMPLE_VERSION_INFO=0.2.3.post2+torch26fancy', '-GNinja', '-DCMAKE_MAKE_PROGRAM:FILEPATH=/opt/anaconda3/envs/kt/bin/ninja']
  Standard output: b'-- The C compiler identification is GNU 11.4.0\n-- The CXX compiler identification is GNU 11.4.0\n-- Detecting C compiler ABI info\n-- Detecting C compiler ABI info - done\n-- Check for working C compiler: /usr/bin/cc - skipped\n-- Detecting C compile features\n-- Detecting C compile features - done\n-- Detecting CXX compiler ABI info\n-- Detecting CXX compiler ABI info - done\n-- Check for working CXX compiler: /usr/bin/c++ - skipped\n-- Detecting CXX compile features\n-- Detecting CXX compile features - done\n-- CMAKE_SYSTEM_PROCESSOR: x86_64\n-- x86 detected\n-- pybind11 v2.14.0 dev1\n-- Found PythonInterp: /opt/anaconda3/envs/kt/bin/python (found suitable version "3.11.11", minimum required is "3.7") \n-- Found PythonLibs: /opt/anaconda3/envs/kt/lib/libpython3.11.so\n-- Performing Test HAS_FLTO\n-- Performing Test HAS_FLTO - Success\n-- Found Git: /usr/bin/git (found version "2.34.1") \n-- Looking for pthread.h\n-- Looking for pthread.h - found\n-- Performing Test CMAKE_HAVE_LIBC_PTHREAD\n-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success\n-- Found Threads: TRUE  \n-- Found OpenMP_C: -fopenmp (found version "4.5") \n-- Found OpenMP_CXX: -fopenmp (found version "4.5") \n-- Found OpenMP: TRUE (found version "4.5")  \n-- OpenMP found\n-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with LLAMA_CCACHE=OFF\n-- CMAKE_SYSTEM_PROCESSOR: x86_64\n-- x86 detected\n-- Found CUDA: /opt/rocm/cuda (found version "dcc version: 25.03.0-0\nclang version 17.0.0\nTarget: x86_64-unknown-linux-gnu\nThread model: posix\nInstalledDir: /opt/dtk-25.04-DCC-2503/llvm/bin\n.dcc version: 25.03.0-0\nclang version 17.0.0\nTarget: x86_64-unknown-linux-gnu\nThread model: posix\nInstalledDir: /opt/dtk-25.04-DCC-2503/llvm/bin\n") \n-- ALL_SOURCES: /root/ktransformers/ktransformers/ktransformers_ext/ext_bindings.cpp;/root/ktransformers/ktransformers/ktransformers_ext/cpu_backend/backend.cpp;/root/ktransformers/ktransformers/ktransformers_ext/cpu_backend/task_queue.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/linear.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/mlp.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/moe.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/shared_mem_buffer.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/flags.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/sgemm.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx512f.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avxvnni.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_fma.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_arm80.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx512f.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avxvnni.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_fma.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_arm80.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_unsupported.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_attn.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_load_dump.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_read_write.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_utils.cpp\n-- NUMA support is disabled\n-- NUMA library not found or user not set USE_NUMA - disabling NUMA support\n-- Configuring done\n-- Generating done\n-- Build files have been written to: /root/ktransformers/ktransformers/ktransformers_ext/build\n'
  Standard error: b'\x1b[33mCMake Warning:\n  Value of CUDA_VERSION contained a newline; truncating\n\n\x1b[0m\n\x1b[33mCMake Warning:\n  Manually-specified variables were not used by the project:\n\n    EXAMPLE_VERSION_INFO\n\n\x1b[0m\n\x1b[33mCMake Warning:\n  Value of CUDA_VERSION contained a newline; truncating\n\n\x1b[0m\n'
  building 'KTransformersOps' extension
  creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip
  creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf
  creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin
  Emitting ninja build file /tmp/tmp7u2_ke9h.build-temp/build.ninja...
  Compiling objects...
  Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
  [1/3] /opt/rocm/bin/hipcc  -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17
  FAILED: /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o
  /opt/rocm/bin/hipcc  -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17
  dcc: error: cannot find ROCm device library for gfx1030; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx1100; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx1101; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx908; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx90a; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  [2/3] /opt/rocm/bin/hipcc  -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17
  FAILED: /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o
  /opt/rocm/bin/hipcc  -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17
  dcc: error: cannot find ROCm device library for gfx1030; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx1100; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx1101; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx908; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  dcc: error: cannot find ROCm device library for gfx90a; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library
  [3/3] c++ -MMD -MF /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/binding.o.d -pthread -B /opt/anaconda3/envs/kt/compiler_compat -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /opt/anaconda3/envs/kt/include -fPIC -O2 -isystem /opt/anaconda3/envs/kt/include -fPIC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/binding.cpp -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/binding.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -O3 -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++17
  ninja: build stopped: subcommand failed.
  Traceback (most recent call last):
    File "/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/utils/cpp_extension.py", line 2209, in _run_ninja_build
      subprocess.run(
    File "/opt/anaconda3/envs/kt/lib/python3.11/subprocess.py", line 571, in run
      raise CalledProcessError(retcode, process.args,
  subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

似乎是编译器坚持要编译gfx908等amdgpu架构代码，但是忽略了本地实际上是gfx928
但是软件检测又正常

(kt) root@lf:~/ktransformers# amdgpu-arch 
gfx928

复现步骤

按照rocm.md操作，正常编译到54/54之后的1/3就会出错

环境信息

硬件 环境
AMD Ryzen Threadripper PRO 7975WX 32-Cores
512G RAM
软件 环境
ubuntu 22.04
anaconda create -n kt python=3.11
conda activate kt
cuda-toolkit-12.8
海光按照官方文档安装驱动
(kt) root@lf:~/ktransformers# hycc -v

HIP_PATH=/opt/dtk-25.04-DCC-2503/hip
HIP_PLATFORM=mix
ROCM_PATH=/opt/rocm
CLANG_PATH=/opt/rocm/llvm/bin
HIP_INCLUDE_PATH=/opt/rocm/include
HIP_LIB_PATH=/opt/dtk-25.04-DCC-2503/hip/lib
DEVICE_LIB_PATH=/opt/rocm/amdgcn/bitcode
hycc args: -v
hycc cmd: /opt/rocm/llvm/bin/clang++  -v
clang version 17.0.0
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /opt/rocm/llvm/bin
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/11
Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/12
Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/12
Candidate multilib: .;@m64
Selected multilib: .;@m64
Found CUDA installation: /usr/local/cuda, version 
Found HIP installation: /opt/rocm, version 6.1.25096
Originally created by @yeungtuzi on GitHub (Mar 27, 2025). Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/984 ### 检查清单 - [x] 1. 我已经搜索过相关问题，但未能获得预期的帮助 - [x] 2. 该问题在最新版本中尚未修复 - [x] 3. 请注意，如果您提交的BUG相关 issue 缺少对应环境信息和最小可复现示例，我们将难以复现和定位问题，降低获得反馈的可能性 - [x] 4. 如果您提出的不是bug而是问题，请在讨论区发起讨论 https://github.com/kvcache-ai/ktransformers/discussions。否则该 issue 将被关闭 - [x] 5. 为方便社区交流，我将使用中文/英文或附上中文/英文翻译（如使用其他语言）。未附带翻译的非中文/英语内容可能会被关闭 ### 问题描述 在make dev_install的时候报错 ``` [54/54] : && /usr/bin/c++ -fPIC -O3 -ffast-math -O3 -DNDEBUG -flto -shared -o /tmp/tmpuqdwl1qf.build-lib/cpuinfer_ext.cpython-311-x86_64-linux-gnu.so CMakeFiles/cpuinfer_ext.dir/ext_bindings.cpp.o CMakeFiles/cpuinfer_ext.dir/cpu_backend/backend.cpp.o CMakeFiles/cpuinfer_ext.dir/cpu_backend/task_queue.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/linear.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/mlp.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/moe.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/llamafile/shared_mem_buffer.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/flags.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/iqk_mul_mat_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/sgemm.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avx512f.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_avxvnni.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_fma.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_arm80.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_mixmul_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx2.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avx512f.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_avxvnni.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_fma.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_amd_zen4.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_arm80.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_sgemm_arm82.cpp.o CMakeFiles/cpuinfer_ext.dir/root/ktransformers/third_party/llamafile/tinyblas_cpu_unsupported.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_attn.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_load_dump.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_read_write.cpp.o CMakeFiles/cpuinfer_ext.dir/operators/kvcache/kvcache_utils.cpp.o -Wl,-rpath,/usr/local/cuda/lib64 third_party/llama.cpp/libllama.a /usr/local/cuda/lib64/libcudart.so /usr/lib/gcc/x86_64-linux-gnu/11/libgomp.so /usr/lib/x86_64-linux-gnu/libpthread.a && cd /root/ktransformers/ktransformers/ktransformers_ext/build && /usr/bin/strip /tmp/tmpuqdwl1qf.build-lib/cpuinfer_ext.cpython-311-x86_64-linux-gnu.so lto-wrapper: warning: using serial compilation of 15 LTRANS jobs CMake args: ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/tmp/tmpuqdwl1qf.build-lib/', '-DPYTHON_EXECUTABLE=/opt/anaconda3/envs/kt/bin/python', '-DCMAKE_BUILD_TYPE=Release', '-DKTRANSFORMERS_USE_CUDA=ON'] CMake args: ['-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/tmp/tmpuqdwl1qf.build-lib/', '-DPYTHON_EXECUTABLE=/opt/anaconda3/envs/kt/bin/python', '-DCMAKE_BUILD_TYPE=Release', '-DKTRANSFORMERS_USE_CUDA=ON', '-DLLAMA_NATIVE=ON', '-DEXAMPLE_VERSION_INFO=0.2.3.post2+torch26fancy', '-GNinja', '-DCMAKE_MAKE_PROGRAM:FILEPATH=/opt/anaconda3/envs/kt/bin/ninja'] Standard output: b'-- The C compiler identification is GNU 11.4.0\n-- The CXX compiler identification is GNU 11.4.0\n-- Detecting C compiler ABI info\n-- Detecting C compiler ABI info - done\n-- Check for working C compiler: /usr/bin/cc - skipped\n-- Detecting C compile features\n-- Detecting C compile features - done\n-- Detecting CXX compiler ABI info\n-- Detecting CXX compiler ABI info - done\n-- Check for working CXX compiler: /usr/bin/c++ - skipped\n-- Detecting CXX compile features\n-- Detecting CXX compile features - done\n-- CMAKE_SYSTEM_PROCESSOR: x86_64\n-- x86 detected\n-- pybind11 v2.14.0 dev1\n-- Found PythonInterp: /opt/anaconda3/envs/kt/bin/python (found suitable version "3.11.11", minimum required is "3.7") \n-- Found PythonLibs: /opt/anaconda3/envs/kt/lib/libpython3.11.so\n-- Performing Test HAS_FLTO\n-- Performing Test HAS_FLTO - Success\n-- Found Git: /usr/bin/git (found version "2.34.1") \n-- Looking for pthread.h\n-- Looking for pthread.h - found\n-- Performing Test CMAKE_HAVE_LIBC_PTHREAD\n-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success\n-- Found Threads: TRUE \n-- Found OpenMP_C: -fopenmp (found version "4.5") \n-- Found OpenMP_CXX: -fopenmp (found version "4.5") \n-- Found OpenMP: TRUE (found version "4.5") \n-- OpenMP found\n-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with LLAMA_CCACHE=OFF\n-- CMAKE_SYSTEM_PROCESSOR: x86_64\n-- x86 detected\n-- Found CUDA: /opt/rocm/cuda (found version "dcc version: 25.03.0-0\nclang version 17.0.0\nTarget: x86_64-unknown-linux-gnu\nThread model: posix\nInstalledDir: /opt/dtk-25.04-DCC-2503/llvm/bin\n.dcc version: 25.03.0-0\nclang version 17.0.0\nTarget: x86_64-unknown-linux-gnu\nThread model: posix\nInstalledDir: /opt/dtk-25.04-DCC-2503/llvm/bin\n") \n-- ALL_SOURCES: /root/ktransformers/ktransformers/ktransformers_ext/ext_bindings.cpp;/root/ktransformers/ktransformers/ktransformers_ext/cpu_backend/backend.cpp;/root/ktransformers/ktransformers/ktransformers_ext/cpu_backend/task_queue.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/linear.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/mlp.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/moe.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/llamafile/shared_mem_buffer.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/flags.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/iqk_mul_mat_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/sgemm.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avx512f.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_avxvnni.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_fma.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_arm80.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_mixmul_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx2.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avx512f.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_avxvnni.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_fma.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_amd_zen4.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_arm80.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_sgemm_arm82.cpp;/root/ktransformers/ktransformers/ktransformers_ext/../../third_party/llamafile/tinyblas_cpu_unsupported.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_attn.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_load_dump.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_read_write.cpp;/root/ktransformers/ktransformers/ktransformers_ext/operators/kvcache/kvcache_utils.cpp\n-- NUMA support is disabled\n-- NUMA library not found or user not set USE_NUMA - disabling NUMA support\n-- Configuring done\n-- Generating done\n-- Build files have been written to: /root/ktransformers/ktransformers/ktransformers_ext/build\n' Standard error: b'\x1b[33mCMake Warning:\n Value of CUDA_VERSION contained a newline; truncating\n\n\x1b[0m\n\x1b[33mCMake Warning:\n Manually-specified variables were not used by the project:\n\n EXAMPLE_VERSION_INFO\n\n\x1b[0m\n\x1b[33mCMake Warning:\n Value of CUDA_VERSION contained a newline; truncating\n\n\x1b[0m\n' building 'KTransformersOps' extension creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf creating /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin Emitting ninja build file /tmp/tmp7u2_ke9h.build-temp/build.ninja... Compiling objects... Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) [1/3] /opt/rocm/bin/hipcc -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17 FAILED: /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o /opt/rocm/bin/hipcc -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/custom_gguf/dequant.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17 dcc: error: cannot find ROCm device library for gfx1030; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx1100; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx1101; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx908; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx90a; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library [2/3] /opt/rocm/bin/hipcc -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17 FAILED: /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o /opt/rocm/bin/hipcc -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.hip -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/gptq_marlin/gptq_marlin.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -DCUDA_HAS_FP16=1 -D__HIP_NO_HALF_OPERATORS__=1 -D__HIP_NO_HALF_CONVERSIONS__=1 -O3 -Xcompiler -fPIC -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 --offload-arch=gfx900 --offload-arch=gfx906 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx942 --offload-arch=gfx1030 --offload-arch=gfx1100 --offload-arch=gfx1101 -fno-gpu-rdc -std=c++17 dcc: error: cannot find ROCm device library for gfx1030; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx1100; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx1101; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx908; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library dcc: error: cannot find ROCm device library for gfx90a; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library [3/3] c++ -MMD -MF /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/binding.o.d -pthread -B /opt/anaconda3/envs/kt/compiler_compat -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /opt/anaconda3/envs/kt/include -fPIC -O2 -isystem /opt/anaconda3/envs/kt/include -fPIC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/torch/csrc/api/include -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/TH -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THC -I/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/include/THH -I/opt/rocm/include -I/opt/anaconda3/envs/kt/include/python3.11 -c -c /root/ktransformers/ktransformers/ktransformers_ext/hip/binding.cpp -o /tmp/tmp7u2_ke9h.build-temp/ktransformers/ktransformers_ext/hip/binding.o -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -fPIC -D__HIP_PLATFORM_AMD__=1 -DUSE_ROCM=1 -DHIPBLAS_V2 -O3 -DKTRANSFORMERS_USE_CUDA -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=KTransformersOps -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++17 ninja: build stopped: subcommand failed. Traceback (most recent call last): File "/opt/anaconda3/envs/kt/lib/python3.11/site-packages/torch/utils/cpp_extension.py", line 2209, in _run_ninja_build subprocess.run( File "/opt/anaconda3/envs/kt/lib/python3.11/subprocess.py", line 571, in run raise CalledProcessError(retcode, process.args, subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1. ``` 似乎是编译器坚持要编译gfx908等amdgpu架构代码，但是忽略了本地实际上是gfx928 但是软件检测又正常 ``` (kt) root@lf:~/ktransformers# amdgpu-arch gfx928 ``` ### 复现步骤 按照rocm.md操作，正常编译到54/54之后的1/3就会出错 ### 环境信息 硬件 环境 AMD Ryzen Threadripper PRO 7975WX 32-Cores 512G RAM 软件 环境 ubuntu 22.04 anaconda create -n kt python=3.11 conda activate kt cuda-toolkit-12.8 海光按照官方文档安装驱动 (kt) root@lf:~/ktransformers# hycc -v ``` HIP_PATH=/opt/dtk-25.04-DCC-2503/hip HIP_PLATFORM=mix ROCM_PATH=/opt/rocm CLANG_PATH=/opt/rocm/llvm/bin HIP_INCLUDE_PATH=/opt/rocm/include HIP_LIB_PATH=/opt/dtk-25.04-DCC-2503/hip/lib DEVICE_LIB_PATH=/opt/rocm/amdgcn/bitcode hycc args: -v hycc cmd: /opt/rocm/llvm/bin/clang++ -v clang version 17.0.0 Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /opt/rocm/llvm/bin Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/11 Found candidate GCC installation: /usr/lib/gcc/x86_64-linux-gnu/12 Selected GCC installation: /usr/lib/gcc/x86_64-linux-gnu/12 Candidate multilib: .;@m64 Selected multilib: .;@m64 Found CUDA installation: /usr/local/cuda, version Found HIP installation: /opt/rocm, version 6.1.25096 ```
bsakaguchi commented 2026-10-29 14:43:01 +00:00
Author
Owner
Copy link

@yeungtuzi commented on GitHub (Mar 27, 2025):

检索代码，发现没有跟gfx928相关的内容

(kt) root@lf:~/ktransformers# grep "gfx9" -R *

ktransformers/ktransformers_ext/cpu_backend/vendors/hip.h:#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__)
ktransformers/ktransformers_ext/cpu_backend/vendors/hip.h:#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__)
ktransformers/ktransformers_ext/vendors/hip.h:#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__)
ktransformers/ktransformers_ext/vendors/hip.h:#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__)
third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile:    gfx900 \
third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile:    gfx906 \
third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile:    gfx908 \
third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile:    gfx90a \
third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile:    gfx900 \
third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile:    gfx906 \
third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile:    gfx908 \
third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile:    gfx90a \
third_party/llama.cpp/.devops/full-rocm.Dockerfile:    gfx900 \
third_party/llama.cpp/.devops/full-rocm.Dockerfile:    gfx906 \
third_party/llama.cpp/.devops/full-rocm.Dockerfile:    gfx908 \
third_party/llama.cpp/.devops/full-rocm.Dockerfile:    gfx90a \
third_party/llama.cpp/ggml-cuda/common.cuh:#if defined(__gfx906__) || defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx1030__)
third_party/llama.cpp/ggml-cuda/common.cuh:#elif defined(__gfx1010__) || defined(__gfx900__)

但是我在海光提供的DAS目录中找到了相关定义
(kt) root@lf:/opt# ls -ld rocm
lrwxrwxrwx 1 root root 18 Mar 27 04:40 rocm -> dtk-25.04-DCC-2503
grep gfx928 -R * rocm/

env.csh:setenv AMDGPU_TARGETS "gfx906;gfx926;gfx928;gfx936"
env.sh:export AMDGPU_TARGETS="gfx906;gfx926;gfx928;gfx936"
hip/bin/rocm_agent_enumerator:echo "gfx928"
grep: hip/bin/hipcc: binary file matches
hip/bin/target.lst:gfx928
grep: hip/lib/hipkernel.bin.gfx928: binary file matches
hip/lib/cmake/hip/hip-config.cmake:  set(AMDGPU_TARGETS "gfx906;gfx926;gfx928;gfx936" CACHE STRING "DCU targets to compile for")
<!-- gh-comment-id:2756700965 --> @yeungtuzi commented on GitHub (Mar 27, 2025): 检索代码，发现没有跟gfx928相关的内容 (kt) root@lf:~/ktransformers# grep "gfx9" -R * ``` ktransformers/ktransformers_ext/cpu_backend/vendors/hip.h:#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) ktransformers/ktransformers_ext/cpu_backend/vendors/hip.h:#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__) ktransformers/ktransformers_ext/vendors/hip.h:#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) ktransformers/ktransformers_ext/vendors/hip.h:#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__) third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile: gfx900 \ third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile: gfx906 \ third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile: gfx908 \ third_party/llama.cpp/.devops/llama-server-rocm.Dockerfile: gfx90a \ third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile: gfx900 \ third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile: gfx906 \ third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile: gfx908 \ third_party/llama.cpp/.devops/llama-cli-rocm.Dockerfile: gfx90a \ third_party/llama.cpp/.devops/full-rocm.Dockerfile: gfx900 \ third_party/llama.cpp/.devops/full-rocm.Dockerfile: gfx906 \ third_party/llama.cpp/.devops/full-rocm.Dockerfile: gfx908 \ third_party/llama.cpp/.devops/full-rocm.Dockerfile: gfx90a \ third_party/llama.cpp/ggml-cuda/common.cuh:#if defined(__gfx906__) || defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx1030__) third_party/llama.cpp/ggml-cuda/common.cuh:#elif defined(__gfx1010__) || defined(__gfx900__) ``` 但是我在海光提供的DAS目录中找到了相关定义 (kt) root@lf:/opt# ls -ld rocm lrwxrwxrwx 1 root root 18 Mar 27 04:40 rocm -> dtk-25.04-DCC-2503 grep gfx928 -R * rocm/ ``` env.csh:setenv AMDGPU_TARGETS "gfx906;gfx926;gfx928;gfx936" env.sh:export AMDGPU_TARGETS="gfx906;gfx926;gfx928;gfx936" hip/bin/rocm_agent_enumerator:echo "gfx928" grep: hip/bin/hipcc: binary file matches hip/bin/target.lst:gfx928 grep: hip/lib/hipkernel.bin.gfx928: binary file matches hip/lib/cmake/hip/hip-config.cmake: set(AMDGPU_TARGETS "gfx906;gfx926;gfx928;gfx936" CACHE STRING "DCU targets to compile for") ```
bsakaguchi commented 2026-10-29 14:43:01 +00:00
Author
Owner
Copy link

@south-ocean commented on GitHub (Mar 27, 2025):

你好，ktransformers 海光代码分支是基于DTK和DAS软件栈编译的，需要安装DAS软件栈配套的AI whl包，尤其是DAS下面的torch和fastpt，然后可以实现ktransformer cuda代码的快速兼容适配。
你可以先确认下环境是否配置好，比如torch、apex、fastpt等是否安装完成并可以正常导入，并且激活dtk下面的cuda env.sh，实现cuda代码兼容。具体编译方式和软件栈资源可见开发者社区https://developer.sourcefind.cn/。
目前可有开箱即用的容器方式使用，见https://forum.sourcefind.cn/thread/980。

Image

<!-- gh-comment-id:2758152772 --> @south-ocean commented on GitHub (Mar 27, 2025): 你好，ktransformers 海光代码分支是基于DTK和DAS软件栈编译的，需要安装DAS软件栈配套的AI whl包，尤其是DAS下面的torch和fastpt，然后可以实现ktransformer cuda代码的快速兼容适配。 你可以先确认下环境是否配置好，比如torch、apex、fastpt等是否安装完成并可以正常导入，并且激活dtk下面的cuda env.sh，实现cuda代码兼容。具体编译方式和软件栈资源可见开发者社区https://developer.sourcefind.cn/。 目前可有开箱即用的容器方式使用，见https://forum.sourcefind.cn/thread/980。 ![Image](https://github.com/user-attachments/assets/872b58f5-6605-47e6-821a-d458bc25fb96)
bsakaguchi commented 2026-10-29 14:43:01 +00:00
Author
Owner
Copy link

@yeungtuzi commented on GitHub (Mar 28, 2025):

编译已经成功，目前试验性跑一下，似乎flashinfer没有rocm/hygon dtk的支持？那么是否意味着后续多并发就没辙了

<!-- gh-comment-id:2760428681 --> @yeungtuzi commented on GitHub (Mar 28, 2025): 编译已经成功，目前试验性跑一下，似乎flashinfer没有rocm/hygon dtk的支持？那么是否意味着后续多并发就没辙了
bsakaguchi commented 2026-10-29 14:43:01 +00:00
Author
Owner
Copy link

@yeungtuzi commented on GitHub (Mar 28, 2025):

已经完成编译并可以运行，有胡言乱语的问题，提交了新的issue，本issue似乎可以关闭了

<!-- gh-comment-id:2760501742 --> @yeungtuzi commented on GitHub (Mar 28, 2025): 已经完成编译并可以运行，有胡言乱语的问题，提交了新的issue，本issue似乎可以关闭了
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
auto/sync-sglang
gh-pages
docs/kt-ft-figure2-single-column-20261010
perf/bf16-grad-norm-20261010
docs/kt-ft-paper-20261008
release/20260911-carrier-repair
release/20260911-ci-python
release/20260911-pypi
ci/four-main-pypi-release
ci/model-e2e-acceptance
release/0.7.0.post3-frozen-pypi
fix/glm53-index-share-sync
release/0.7.0.post2-frozen-pypi
hotfix/0.7.0.post2-pypi-runtime
fix/avx2-wheel-feature-isolation
fix/release-self-hosted-runner
release/automate-kt-sft-pypi
support-glm5-next-djw
docs/deepseek-v4-flash-docker
ovowei-patch-3
support-glm52-djw
doc-reorg-entrypoints-support-matrix
bump-sglang-pr31
docs-v061-refresh
fp4-moe-amx
flatten-ktransformers-shim
archive-kt-sft
kt-kernel-torch-range-v061
sft-whl
sft
docs/gosim-2026-and-roadmap-link
ovowei-patch-2
ovowei-patch-1
revert-1918-fix/rocm-cpuinfer-stream-bridge
chore/kt-layerwise-prefill-main
chore/kt-layerwise-prefill-label
sft-rel
support-glm51-djw
release/v0.5.3
chore/kt-run-numa-nodes-unified
mtp
ksft-sglang
avx2
fix/sglang-kt-detection
layerwise-prefill-opt
feat/mistral-loader-support
fix-docker-build
minimax-m2.5
qwen3.5
add-maintainers
oql-doc
cli-dev-ljq
ljq-cli-dev-new
JimmyPeilinLi-patch-2
kimi_k2.5
oql-dev
experts_sched
JimmyPeilinLi-patch-1
ErvinXie-patch-1
ksft-llama
glm47_fp8
debug_llama
refactor_sft
ljq-cli-dev
refactor_sft_int4_debug
lut-nvfp4
fix-import
fix-deploy-djw
cpu-fix
ljq-dev-1
route-moe
sosp25-ae
v0.7.2
paper-ktft-v1.1
paper-ktft-v1
v0.7.1
v0.7.0.post4
v0.7.0.post1
v0.7.0
v0.6.4
v0.6.3.post1
v0.6.3
v0.6.2.post3
v0.6.2.post2
v0.6.2.post1
v0.6.2
v0.6.1
v0.6.1.post1
v0.5.3
v0.5.2.post1
v0.5.1
v0.5.0.post1
v0.5.0
v0.4.4
v0.4.3
v0.4.2
v0.4.1
v0.3.2
v0.3.1
v0.3
v0.2.4post1
v0.2.4
v0.2.3post2
v0.2.3post1
v0.2.3
v0.2.2
v0.2.2rc2
v0.2.2rc1
v0.2.1.post1
v0.2.1
v0.2.0
v0.1.4
v0.1.3
v0.1.2
v0.1.1
v0.1.0
Labels
Clear labels   
bug

   
duplicate

   
enhancement

   
good first issue

   
help wanted

   
pull-request
Mirrored from GitHub Pull Request

  
stale
No labels
bug
duplicate
enhancement
good first issue
help wanted
pull-request
stale
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
kvcache-ai/ktransformers#698
No description provided.