Bring SYCL* to Pre-exascale Supercomputing with DPC++ for CUDA*

Perlmutter is a pre-exascale supercomputer for NERSC at Lawrence Berkeley National Laboratory that consists of more than 6000 NVIDIA* A100 GPUs. Codeplay is working in association with NERSC, LBNL, and Argonne National Laboratory to enable developers to use this supercomputer and ALCF ThetaGPU machine to write highly accelerated applications using the SYCL* programming model.

This session demonstrates how this open-source work benefits the whole community and enables developers to bring their CUDA* code to SYCL while still being able to obtain comparable performance on NVIDIA* GPUs.

Speaker

Steffen Larsen is a staff software engineer at Codeplay* and has been working on DPC++ for CUDA since early in its development. With SYCL 2020 released and interest for SYCL on the rise, Steffen and his team at Codeplay are working to bring SYCL 2020 feature support to DPC++ for CUDA as they are implemented in DPC++.