The ALCF is hosting a two-part workshop for Argonne's upcoming exascale system, Aurora.
This is an invitation-only workshop for researchers participating in the ALCF's Aurora Early Science Program and DOE's Exascale Computing Project. Please note that your institution must have an CNDA agreement with Intel and you must be listed on our Multiparty Authorization Agreement (MPA) for you to be allowed to attend the workshop.
Part One (Oct. 26-27)
Aurora system update:
- Overview and updates on Aurora hardware
- Overview and updates on Aurora software
Aurora applications:
- Survey of approaches to Aurora readiness
- Performance and portability across multiple GPU platforms
- Preparing applications to use multiple GPUs on a node
Performance evaluation tutorials:
- Using Intel VTune
- Using Intel Advisor
Programming models and frameworks tracks:
- Track 1: OpenMP Target Offload
- Track 2: DPC++ and SYCL
- Track 3: AI frameworks and software
Part Two (Dec. 7-9)
- Hands-on with Aurora testbed hardware
- Highlights of recent progress with applications