Benefits of Multi-rail Cluster Architectures for GPU-based Nodes

Craig Tierney is a Solutions Architect at NVIDIA

In this video from the MVAPICH User Group meeting, Craig Tierney from NVIDIA presents: Benefits of Multi-rail Cluster Architectures for GPU-based Nodes.

“As high performance computing moves toward GPU-accelerated architectures, single node application performance can be between 3x and 75x faster than the CPUs alone. Performance increases of this size will require increases in network bandwidth and message rate to prevent the network from becoming the bottleneck in scalability. In this talk, we will present results from NVLink enabled systems connected via quad-rail EDR Infiniband.”

Craig Tierney is a Solutions Architect in the Worldwide Field Organization at NVIDIA. His primary roles are to support deep learning and high performance computing. He is focused on applying deep learning to weather and climate research and leveraging the power of typical HPC technologies to accelerate deep learning. Prior to joining NVIDIA, Craig spent more than a decade providing high performance computing architecture and computational science support to NOAA and several other government and educational organizations including DOE, NASA, DOD and Stanford University. Craig holds a Ph.D. in Aerospace Engineering Sciences from the University of Colorado at Boulder.

See more talks from the MVAPICH User Group

Sign up for our insideHPC Newsletter