Resources

The Aurora exascale system is currently being deployed at Argonne National Lab. The system, utilizing Intel’s new Data Center Max Series GPUs (a.k.a. PVC) and Xeon Max Series CPU with HBM, will provide a uniquely powerful platform for leading-edge HPC, AI, and data-intensive computing applications. Scientists at Argonne National Laboratory, in collaboration with the Exascale Computing Project, Intel, and several other institutions, are preparing several dozen applications and workflows to run at scale on the Aurora system. This talk will present an overview of the Aurora system and highlights from the experience of preparing applications for the system. In addition, promising early performance results on the Aurora hardware will be shown.

Keyword(s): Exascale,Aurora,PVC,Ponte Vecchio,GPU Max,CPU Max,HBM,Data Center Max Series GPUs,Xeon Max Series CPU with HBM

Mapping Cores, CHAs, and Addresses in the Xeon Platinum 8380 by John McCalpin, Texas Advanced Computing Center, University of Texas at Austin

Keyword(s): Xeon Platinum 8380,Mapping Intel Xeon Processors,Ice Lake,conventions for numbering CHAs and cores

Improving MPI+Threads with MPIX_Stream by Ken Raffenetti, Argonne National Laboratory

Keyword(s): MPI+Threads,MPIX_Stream,MPI Asynchronous Progress,Virtual Communication Interface (VCI),Stream Communicator,MPI Progress Model

Keynote: New Era for Intel HPC Acceleration: Architecture, Systems & Software by Hong Jiang, Intel Corporation

Keyword(s): Exascale Compute Platform,oneAPI Software Stack,Ponte Vecchio,Application Performance,Zettascale

Keynote on the future of HPC and scientific discovery by Rick Stevens, Argonne National Laboratory

Keyword(s): Exascale,Science,Machine Learning,CPU,GPU,LLMs,Inverse Design,DeepFix

Recovery of Distributed Iterative Solvers for Linear Systems with Non-Volatile RAM by Yehonatan Fridman, Israel Atomic Energy Commission and Ben-Gurion University

Keyword(s): Exascale,Non-Volatile RAM (NVRAM),Persistent Memory,Linear Iterative Solvers,Exact State Reconstruction (ESR),Preconditioned Conjugate Gradient (PCG)

Best Practices for Benchmarking Diverse Architectures with Varied Workloads (agenda slide)

Keyword(s): IXPUG BoF at SC21 agenda slide

IXPUG Overview and Activities (IXPUG BoF at SC21)

Keyword(s): IXPUG overview and activities

A Simple Cost-model for Comparing Diverse Architectures

Keyword(s): cost model,diverse hardware,time-to-solution,memory footprint,result depreciation,compute cost,cost-per-run,Amdahl’s law

Best Practices for Benchmarking Diverse Architectures with Varied Workloads

Keyword(s): benchmarks,performance,double precision performance,Roofline Model,benchmark parameters,reproducibility,Continuous Benchmarking,Continuous Integration