-
Notifications
You must be signed in to change notification settings - Fork 448
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.
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
Failed to compile heavyDB; CUDA architecture not detected #822
Comments
To compile the GPU version, you need CUDA 11.8; version 11.4 is outdated. Additionally, ensure your GPU driver is at least version 520.x (though it's not required at compile time). If you encounter issues with CMake, make sure to run mapd-deps.sh found in /usr/local/mapd-deps. This directory includes the Thrift library and other dependencies not present in the standard Ubuntu repositories. |
Hi, I upgraded the CUDA version to 12.1 and GPU driver to 530.x. However the cmake is still failing. Do I need to make any changes in the CMakeLists.txt file? Thanks for the suggestion though! |
Hello, Just wanted to let you know that you don't need to make any modifications to the CMakeLists.txt file. Instead, you can try adding "-DCMAKE_CUDA_ARCHITECTURES=75" (replace 75 with the appropriate version for your setup) to your command. However, please note that you may encounter an error if you run this command with an unsupported CUDA version or if you have already run the config script. To avoid this, run the command "source /usr/local/mapd-deps/mapd-deps.sh" before running the cmake command. I haven't any issue using cuda 12.1 on my system. could you share the entire output of your cmake command? mine is:
|
Hi,
I am unable to complie heavyDB. I have both the nvidia compiler and nvidia drivers setup in the system.
I have followed the steps below --
./scripts/mapd-deps-prebuilt.sh && mkdir build && cd build && cmake ..
The scripts/mapd-deps-prebuild.sh executes successfully.
The error code received is --
CMake Error at /usr/local/share/cmake-3.25/Modules/CMakeDetermineCUDACompiler.cmake:603 (message): Failed to detect a default CUDA architecture. Compiler output: Call Stack (most recent call first): CMakeLists.txt:194 (enable_language)
My system environment is as follows:
Let me know if I need to downgrade/upgrade any software to allow the build to happen as well.
Alternatively, when I try to compile by disabling CUDA like --
cmake -DENABLE_CUDA=off ..
I get the following error:
The text was updated successfully, but these errors were encountered: