There is a newer version of cuDNNThe NVIDIA CUDA Deep Neural Network library (cuDNN) is a GPU-accelerated library of primitives for deep neural networks. cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. cuDNN is part of the NVIDIA Deep Learning SDK.
Accessing cuDNN 18.104.22.168-CUDA-11.3.1
To load the module for cuDNN 22.214.171.124-CUDA-11.3.1 please use this command on the BEAR systems (BlueBEAR, BEARCloud VMs, and CaStLeS VMs):
module load cuDNN/126.96.36.199-CUDA-11.3.1
BEAR Apps Version
For more information visit the cuDNN website.
This version of cuDNN has a direct dependency on: CUDA/11.3.1
This version of cuDNN is a direct dependent of: AlphaFold/2.1.1-foss-2021a-CUDA-11.3.1 AlphaFold/2.2.2-foss-2021a-CUDA-11.3.1 jax/0.3.9-foss-2021a-CUDA-11.3.1 jax/0.2.24-foss-2021a-CUDA-11.3.1 OpenCV/4.5.3-foss-2021a-CUDA-11.3.1-contrib PyTorch/1.9.1-foss-2021a-CUDA-11.3.1-imkl PyTorch/1.10.0-foss-2021a-CUDA-11.3.1 TensorFlow/2.6.0-foss-2021a-CUDA-11.3.1
These versions of cuDNN are available on the BEAR systems (BlueBEAR, BEARCloud VMs, and CaStLeS VMs). These will be retained in accordance with our Applications Support and Retention Policy.
Last modified on 11th October 2021