Hello!
I have a Nvidia GTX 760 I am trying to use with PyTorch in Arch Linux. Searching online it appears I need to build from source. I have CUDA 10.2.89 installed with driver version 440.82.
If anyone has advice on how to fix this issue it would be appreciated.
[1656/2689] [1649/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_linear_op.cu.o
FAILED: caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_linear_op.cu.o
cd /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators && /home/admin/.conda/envs/torch_gpu/bin/cmake -E make_directory /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/. && /home/admin/.conda/envs/torch_gpu/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING=Release -D generated_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_linear_op.cu.o -D generated_cubin_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_linear_op.cu.o.cubin.txt -P /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_linear_op.cu.o.Release.cmake
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(164): error: expression must have a constant value
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(90): note: expression cannot be interpreted
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(110): note: called from:
detected during:
instantiation of "c10::string_view c10::util::get_fully_qualified_type_name<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/typeid.h(415): here
instantiation of "c10::string_view caffe2::TypeMeta::TypeName<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/caffe2/core/blob.h(88): here
1 error detected in the compilation of "/tmp/tmpxft_00007942_00000000-4_elementwise_linear_op.cpp4.ii".
CMake Error at torch_cuda_generated_elementwise_linear_op.cu.o.Release.cmake:281 (message):
Error generating file
/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_linear_op.cu.o
[1650/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_enforce_finite_op.cu.o
FAILED: caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_enforce_finite_op.cu.o
cd /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators && /home/admin/.conda/envs/torch_gpu/bin/cmake -E make_directory /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/. && /home/admin/.conda/envs/torch_gpu/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING=Release -D generated_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_enforce_finite_op.cu.o -D generated_cubin_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_enforce_finite_op.cu.o.cubin.txt -P /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_enforce_finite_op.cu.o.Release.cmake
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(164): error: expression must have a constant value
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(90): note: expression cannot be interpreted
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(110): note: called from:
detected during:
instantiation of "c10::string_view c10::util::get_fully_qualified_type_name<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/typeid.h(415): here
instantiation of "c10::string_view caffe2::TypeMeta::TypeName<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/caffe2/core/blob.h(88): here
1 error detected in the compilation of "/tmp/tmpxft_00007938_00000000-4_enforce_finite_op.cpp4.ii".
CMake Error at torch_cuda_generated_enforce_finite_op.cu.o.Release.cmake:281 (message):
Error generating file
/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_enforce_finite_op.cu.o
[1651/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_filler_op.cu.o
FAILED: caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_filler_op.cu.o
cd /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators && /home/admin/.conda/envs/torch_gpu/bin/cmake -E make_directory /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/. && /home/admin/.conda/envs/torch_gpu/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING=Release -D generated_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_filler_op.cu.o -D generated_cubin_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_filler_op.cu.o.cubin.txt -P /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_filler_op.cu.o.Release.cmake
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(164): error: expression must have a constant value
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(90): note: expression cannot be interpreted
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(110): note: called from:
detected during:
instantiation of "c10::string_view c10::util::get_fully_qualified_type_name<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/typeid.h(415): here
instantiation of "c10::string_view caffe2::TypeMeta::TypeName<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/caffe2/core/blob.h(88): here
1 error detected in the compilation of "/tmp/tmpxft_00007948_00000000-4_filler_op.cpp4.ii".
CMake Error at torch_cuda_generated_filler_op.cu.o.Release.cmake:281 (message):
Error generating file
/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_filler_op.cu.o
[1652/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_cross_entropy_op.cu.o
FAILED: caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_cross_entropy_op.cu.o
cd /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators && /home/admin/.conda/envs/torch_gpu/bin/cmake -E make_directory /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/. && /home/admin/.conda/envs/torch_gpu/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING=Release -D generated_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_cross_entropy_op.cu.o -D generated_cubin_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_cross_entropy_op.cu.o.cubin.txt -P /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_cross_entropy_op.cu.o.Release.cmake
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(164): error: expression must have a constant value
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(90): note: expression cannot be interpreted
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(110): note: called from:
detected during:
instantiation of "c10::string_view c10::util::get_fully_qualified_type_name<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/typeid.h(415): here
instantiation of "c10::string_view caffe2::TypeMeta::TypeName<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/caffe2/core/blob.h(88): here
1 error detected in the compilation of "/tmp/tmpxft_00007950_00000000-4_cross_entropy_op.cpp4.ii".
CMake Error at torch_cuda_generated_cross_entropy_op.cu.o.Release.cmake:281 (message):
Error generating file
/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_cross_entropy_op.cu.o
[1653/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_mul_op.cu.o
FAILED: caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_mul_op.cu.o
cd /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators && /home/admin/.conda/envs/torch_gpu/bin/cmake -E make_directory /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/. && /home/admin/.conda/envs/torch_gpu/bin/cmake -D verbose:BOOL=OFF -D build_configuration:STRING=Release -D generated_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_mul_op.cu.o -D generated_cubin_file:STRING=/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_mul_op.cu.o.cubin.txt -P /home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/torch_cuda_generated_elementwise_mul_op.cu.o.Release.cmake
/home/admin/.conda/envs/torch_gpu/pytorch/cmake/../third_party/eigen/Eigen/src/Core/arch/GPU/PacketMathHalf.h(149): warning: missing return statement at end of non-void function "Eigen::internal::ptrue(const Packet &) [with Packet=half2]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(164): error: expression must have a constant value
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(90): note: expression cannot be interpreted
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/TypeIndex.h(110): note: called from:
detected during:
instantiation of "c10::string_view c10::util::get_fully_qualified_type_name<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/c10/util/typeid.h(415): here
instantiation of "c10::string_view caffe2::TypeMeta::TypeName<T>() [with T=caffe2::Tensor]"
/home/admin/.conda/envs/torch_gpu/pytorch/caffe2/core/blob.h(88): here
1 error detected in the compilation of "/tmp/tmpxft_00007973_00000000-4_elementwise_mul_op.cpp4.ii".
CMake Error at torch_cuda_generated_elementwise_mul_op.cu.o.Release.cmake:281 (message):
Error generating file
/home/admin/.conda/envs/torch_gpu/pytorch/build/caffe2/CMakeFiles/torch_cuda.dir/operators/./torch_cuda_generated_elementwise_mul_op.cu.o
[1656/2689] Building NVCC (Device) object caffe2/CMakeFiles/torch_cuda.dir/__/aten/src/THC/torch_cuda_generated_THCTensorMathPairwise.cu.o
ninja: build stopped: subcommand failed.
Traceback (most recent call last):
File "setup.py", line 738, in <module>
build_deps()
File "setup.py", line 315, in build_deps
build_caffe2(version=version,
File "/home/admin/.conda/envs/torch_gpu/pytorch/tools/build_pytorch_libs.py", line 62, in build_caffe2
cmake.build(my_env)
File "/home/admin/.conda/envs/torch_gpu/pytorch/tools/setup_helpers/cmake.py", line 340, in build
self.run(build_args, my_env)
File "/home/admin/.conda/envs/torch_gpu/pytorch/tools/setup_helpers/cmake.py", line 141, in run
check_call(command, cwd=self.build_dir, env=env)
File "/home/admin/.conda/envs/torch_gpu/lib/python3.8/subprocess.py", line 364, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['cmake', '--build', '.', '--target', 'install', '--config', 'Release', '--', '-j', '8']' returned non-zero exit status 1.