GPUs and CUDA
There are GPUs available for general use on the YCRC clusters. In order to use them, you must request them for your job. See the Grace, Farnam, and Milgram pages for hardware and partition specifics. Please do not use nodes with GPUs unless your application or job can make use of them. Any jobs submitted to a GPU partition without having requested a GPU may be terminated without warning.
Monitor Activity and Drivers
The CUDA libraries you load will allow you to compile code against them. To run CUDA-enabled code you must also be running on a node with a gpu allocated and a compatible driver installed. The minimum driver versions are listed on this nvidia developer site.
You can check the available GPUs, their current usage, installed version of the nvidia drivers, and more with the command nvidia-smi. Either in an interactive job or after connecting to a node running your job with ssh, nvidia-smi output should look something like this:
[user@gpu01 ~]$ nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03 Driver Version: 460.32.03 CUDA Version: 11.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... On | 00000000:02:00.0 Off | N/A |
| 23% 34C P8 9W / 250W | 1MiB / 11178MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
Here we see that the node gpu01 is running driver version 460.32.03 and is compatible with CUDA version 11.2. There are no processes using the GPU allocated to this job.
Software
CUDA and cuDNN modules
We have seen varying degrees of success in using the runtime CUDA and cuDNN libraries supplied by various conda channels. If that works for you there may be no need to load additional modules. If not, find the corresponding CUDA and cuDNN combination for your desired environment and load or request those modules. To list all the CUDA and cuDNN modules available:
module avail cuda/
module avail cudnn/
Tensorflow
You can find hints about the correct version of Tensorflow from their tested build configurations. You can also test your install with a simple script that imports Tensorflow (run on a GPU node). If you an ImportError that mentions missing libraries like libcublas.so.9.0, for example, that means that Tensorflow is probably expecting CUDA v 9.0 but cannot find it.
Create an Example Tensorflow-GPU Environment
To create a conda environment without the system CUDA:
module load miniconda
conda create --name tf-condacuda python numpy pandas matplotlib jupyter cudatoolkit=11.2 tensorflow-gpu
To create a conda environment with Tensorflow and uses the module CUDA:
# load modules, including the system CUDA and cuDNN
module load miniconda CUDAcore/11.2.2 cuDNN/8.1.1.33-CUDA-11.2.2
# save module collection for furture use
module save cuda11
#create environment with required dependencies
conda create --name tf-modulecuda python numpy pandas matplotlib jupyter -c conda-forge
# use pip to install tensorflow-gpu
pip install tensorflow-gpu
Use Your Environment
To re-enter your environment you only need the following:
module load miniconda
conda activate tf-condacuda
module restore cuda11
conda activate tf-modulecuda
PyTorch
As with Tensorflow, sometimes the conda-supplied CUDA libraries are sufficient for the version of PyTorch you are installing. If not make sure you have the version of CUDA referenced on the PyTorch site in their install instructions. They also provide instructions on installing previous versions compatible with older versions of CUDA.
Following the instructions on their site, create a PyTorch environment using conda:
module load miniconda
conda create --name pytorch_env pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch
Compile .c or .cpp Files with CUDA code
By default, nvcc expects that host code is in files with a .c or .cpp extension, and device code is in files with a .cu extension. When you mix device code in a .c or .cpp file with host code, the device code will not be recoganized by nvcc unless you add this flag: -x cu.
nvcc -x cu mycuda.cpp -o mycuda.exe