Skip to content

Instantly share code, notes, and snippets.

@jammm

jammm/gist:124b32e652840f5f8879220673a36e1f

Created September 5, 2025 17:34
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save jammm/124b32e652840f5f8879220673a36e1f to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save jammm/124b32e652840f5f8879220673a36e1f to your computer and use it in GitHub Desktop.
Download ZIP
Raw
gistfile1.txt
[557/563] Linking CXX shared library bin\torch_hip.dll
FAILED: [code=4294967295] bin/torch_hip.dll lib/torch_hip.lib
C:\WINDOWS\system32\cmd.exe /C "cd . && D:\jam\venv\Lib\site-packages\cmake\data\bin\cmake.exe -E vs_link_dll --msvc-ver=1944 --intdir=caffe2\CMakeFiles\torch_hip.dir --rc=C:\PROGRA~2\WI3CF2~1\10\bin\100261~1.0\x64\rc.exe --mt=C:\PROGRA~1\MICROS~1\2022\COMMUN~1\VC\Tools\Llvm\x64\bin\llvm-mt.exe --manifests -- D:\jam\venv\Lib\site-packages\_rocm_sdk_devel\lib\llvm\bin\lld-link.exe /nologo @CMakeFiles\torch_hip.rsp /out:bin\torch_hip.dll /implib:lib\torch_hip.lib /pdb:bin\torch_hip.pdb /dll /version:0.0 /machine:x64 /ignore:4049 /ignore:4217 /ignore:4099 /INCREMENTAL:NO && cd ."
LINK: command "D:\jam\venv\Lib\site-packages\_rocm_sdk_devel\lib\llvm\bin\lld-link.exe /nologo @CMakeFiles\torch_hip.rsp /out:bin\torch_hip.dll /implib:lib\torch_hip.lib /pdb:bin\torch_hip.pdb /dll /version:0.0 /machine:x64 /ignore:4049 /ignore:4217 /ignore:4099 /INCREMENTAL:NO /MANIFEST:EMBED,ID=2" failed (exit code 1) with the following output:
lld-link: error: undefined symbol: enum hipError_t __cdecl aotriton::v2::flash::check_gpu(class aotriton::StreamTemplate<struct ihipStream_t *>)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class c10::SymInt, class c10::SymInt> __cdecl at::native::_efficient_attention_forward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<__int64>, class std::optional<__int64>, double, __int64, bool, class std::optional<double>, class std::optional<class at::Tensor> const &, class std::optional<__int64>))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention_backward.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl at::native::_efficient_attention_backward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, __int64, __int64, class at::Tensor const &, double, class at::Tensor const &, class at::Tensor const &, __int64, bool, class std::optional<double>, class std::optional<__int64>, class std::optional<__int64>, bool))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_fwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, float, float, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class std::optional<struct at::Generator> const &))
>>> referenced 5 more times
lld-link: error: undefined symbol: public: __cdecl aotriton::v3::flash::attn_fwd_params::attn_fwd_params(void)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class c10::SymInt, class c10::SymInt> __cdecl at::native::_efficient_attention_forward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<__int64>, class std::optional<__int64>, double, __int64, bool, class std::optional<double>, class std::optional<class at::Tensor> const &, class std::optional<__int64>))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_fwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, float, float, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class std::optional<struct at::Generator> const &))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_varlen_fwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, int, int, float, float, bool, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class std::optional<struct at::Generator> const &))
lld-link: error: undefined symbol: enum hipError_t __cdecl aotriton::v3::flash::attn_fwd(struct aotriton::v3::flash::attn_fwd_params const &, int, class aotriton::StreamTemplate<struct ihipStream_t *>, struct aotriton::v3::flash::attn_options const *)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class c10::SymInt, class c10::SymInt> __cdecl at::native::_efficient_attention_forward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, class std::optional<__int64>, class std::optional<__int64>, double, __int64, bool, class std::optional<double>, class std::optional<class at::Tensor> const &, class std::optional<__int64>))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_fwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, float, float, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class std::optional<struct at::Generator> const &))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_varlen_fwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, int, int, float, float, bool, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class std::optional<struct at::Generator> const &))
lld-link: error: undefined symbol: enum hipError_t __cdecl aotriton::v2::flash::debug_simulate_encoded_softmax(class aotriton::TensorView<4>, float, class aotriton::TensorView<0>, class aotriton::TensorView<0>, unsigned __int64, class aotriton::StreamTemplate<struct ihipStream_t *>)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention.hip.obj:(class at::Tensor & __cdecl at::native::_fill_mem_eff_dropout_mask_(class at::Tensor &, double, __int64, __int64))
lld-link: error: undefined symbol: public: __cdecl aotriton::v3::flash::attn_bwd_params::attn_bwd_params(void)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention_backward.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl at::native::_efficient_attention_backward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, __int64, __int64, class at::Tensor const &, double, class at::Tensor const &, class at::Tensor const &, __int64, bool, class std::optional<double>, class std::optional<__int64>, class std::optional<__int64>, bool))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_bwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, float, float, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class at::Tensor const &, class at::Tensor const &))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_varlen_bwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, int, int, float, float, bool, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class at::Tensor const &, class at::Tensor const &))
lld-link: error: undefined symbol: enum hipError_t __cdecl aotriton::v3::flash::attn_bwd(struct aotriton::v3::flash::attn_bwd_params const &, int, class aotriton::StreamTemplate<struct ihipStream_t *>, struct aotriton::v3::flash::attn_options const *)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/torch_hip_generated_attention_backward.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl at::native::_efficient_attention_backward(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class at::Tensor const &, class std::optional<class at::Tensor> const &, class std::optional<class at::Tensor> const &, __int64, __int64, class at::Tensor const &, double, class at::Tensor const &, class at::Tensor const &, __int64, bool, class std::optional<double>, class std::optional<__int64>, class std::optional<__int64>, bool))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_bwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, float, float, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class at::Tensor const &, class at::Tensor const &))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/flash_attn/aot/torch_hip_generated_mha_all_aot.hip.obj:(class std::tuple<class at::Tensor, class at::Tensor, class at::Tensor, class at::Tensor> __cdecl pytorch_flash::mha_varlen_bwd_aot(class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class std::optional<class at::Tensor> &, class at::Tensor const &, class at::Tensor const &, class std::optional<class at::Tensor> &, int, int, float, float, bool, bool, class std::optional<__int64>, class std::optional<__int64>, bool, class at::Tensor const &, class at::Tensor const &))
lld-link: error: undefined symbol: bool __cdecl aotriton::isArchExperimentallySupported(struct ihipStream_t *)
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/sdp_utils.cpp.obj:(bool __cdecl sdp::`anonymous namespace'::check_flash_attention_hardware_support(struct 0xB543D52::sdp_params const &, bool))
>>> referenced by caffe2/CMakeFiles/torch_hip.dir/__/aten/src/ATen/native/transformers/hip/sdp_utils.cpp.obj:(bool __cdecl sdp::`anonymous namespace'::check_mem_efficient_hardware_support(struct 0xB543D52::sdp_params const &, bool))
ninja: build stopped: subcommand failed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment