NCCL 2.8.3-GCCcore-10.2.0-CUDA-11.1.1
There is a newer install of NCCL
The NVIDIA Collective Communications Library (NCCL) implements multi-GPU and multi-node collective communication primitives that are performance optimized for NVIDIA GPUs.Accessing NCCL 2.8.3-GCCcore-10.2.0-CUDA-11.1.1
To load the module for NCCL 2.8.3-GCCcore-10.2.0-CUDA-11.1.1 please use this command on Baskerville:
📋
module load bask-apps/test
module load NCCL/2.8.3-GCCcore-10.2.0-CUDA-11.1.1
Bask Apps Env
More Information
For more information visit the NCCL website.
Dependencies
This version of NCCL has a direct dependency on: CUDAcore/11.1.1 GCCcore/10.2.0
Required By
This version of NCCL is a direct dependent of: CuPy/8.5.0-fosscuda-2020b Horovod/0.22.1-fosscuda-2020b-PyTorch-1.8.1 libgpuarray/0.7.6-fosscuda-2020b PyTorch/1.8.1-fosscuda-2020b TensorFlow/2.4.1-fosscuda-2020b TensorFlow/2.5.0-fosscuda-2020b
Other Versions
These versions of NCCL are available on Baskerville. These will be retained in accordance with our Applications Support and Retention Policy.
Last modified on 10th July 2021