Hi all,
my computer has a GeForceGT710 graphics card (OS Linux Mint). My understanding is that I need to install pytorch from source code to avoid the “CUDA error: no kernel image is available for execution on this device” warning.
I have so far followed the instructions specified on github (GitHub - pytorch/pytorch: Tensors and Dynamic neural networks in Python with strong GPU acceleration). However, I am getting an error while compiling:
CMake Warning at modules/observers/CMakeLists.txt:12 (add_library):
Cannot generate a safe runtime search path for target caffe2_observers
because files in some directories may conflict with libraries in implicit
directories:runtime library [libcudart.so.10.1] in /usr/lib/x86_64-linux-gnu may be hidden by files in: /home/mareike/miniconda3/envs/ginjinn/lib runtime library [libnvToolsExt.so.1] in /usr/lib/x86_64-linux-gnu may be hidden by files in: /home/mareike/miniconda3/envs/ginjinn/lib runtime library [libcufft.so.10] in /usr/lib/x86_64-linux-gnu may be hidden by files in: /home/mareike/miniconda3/envs/ginjinn/lib runtime library [libcurand.so.10] in /usr/lib/x86_64-linux-gnu may be hidden by files in: /home/mareike/miniconda3/envs/ginjinn/lib runtime library [libcublas.so.10] in /usr/lib/x86_64-linux-gnu may be hidden by files in: /home/mareike/miniconda3/envs/ginjinn/lib
Some of these libraries may not be found correctly.
CMake Error at modules/observers/CMakeLists.txt:12 (add_library):
The install of the caffe2_observers target requires changing an RPATH from
the build tree, but this is not supported with the Ninja generator unless
on an ELF-based platform. The CMAKE_BUILD_WITH_INSTALL_RPATH variable may
be set to avoid this relinking step.CMake Error at modules/observers/CMakeLists.txt:12 (add_library):
The install of the caffe2_observers target requires changing an RPATH from
the build tree, but this is not supported with the Ninja generator unless
on an ELF-based platform. The CMAKE_BUILD_WITH_INSTALL_RPATH variable may
be set to avoid this relinking step.– Generating done
CMake Generate step failed. Build files cannot be regenerated correctly.
My best idea so far is to set CMAKE_BUILD_WITH_INSTALL_RPATH to TRUE. However, I am not too knowledgeable about programming, let alone cmake.
Other information that may be of use: I run gcc 9.3.0, g++ 9.3.0, cudatoolkit 10.1.243, and cudnn 7.6.5. I have everything including dependencies (numpy,… again as specified on github) installed in a conda environment. However, the error also occurs when I install everything globally.
Can someone help me to resolve this?
Best,
Mareike