Aurora Introduction

Help Desk

Hours: 9:00am-5:00pm CT M-F
Email: support@alcf.anl.gov

Aurora

Aurora, Argonne National Laboratory’s first exascale computer is due to arrive in 2021. Scientists will use the new machine to pursue science and engineering breakthroughs beyond what is possible with today’s supercomputers. Aurora, designed in collaboration with Intel and Cray, will support machine learning and data science workloads alongside traditional modeling and simulation workloads.

For decades, the DOE has been building an aggressive scientific computing research program to give our nation a strategic competitive advantage in the advancement of science and technology. Today, the U.S. program is unrivaled in the world and provides groundbreaking discoveries in all fields of inquiry. The massive machines at the core of the DOE supercomputing program are new ground-breaking designs, and each one ten to a hundred times more powerful than other systems used for scientific research

Aurora will feature several technological innovations, and will be built on a future generation of Intel® Xeon® Scalable processor accelerated by Intel’s Xe compute architecture. Cray Slingshot™ fabric and Shasta™ platform will form the backbone of the system. Aurora will also including a revolutionary I/O system—the Distributed Asynchronous Object Store (DAOS)—to support new types of workloads. Programming techniques already in use on current systems, such as OpenMP, will apply directly to Aurora along with newer models including Intel’s DPC++.

System Specifications

Sustained Performance ≥ 1 Exaflop
Aggregate System Memory > 10 PB
High-Performance Storage DAOS: ≥ 230 PB, ≥ 25 TB/s
Lustre: ≥ 150 PB, ≥ 1 TB/s
Compute Node Architecture 2 Intel Xeon scalable processors (Sapphire Rapids),
6 Xe arch-based GP-GPUs (Ponte Vecchio);
Unified memory architecture across CPU & GPU;
All-to-all connectivity within node
CPU-GPU Interconnect PCle
System Interconnect Cray Slingshot; Dragonfly; 8 fabric endpoints per node
Network Switch 25.6 Tb/s per switch, from 64–200 Gbs ports (25 GB/s per direction)
Programming Models OpenMP, SYCL/DPC++, OpenCL, Kokkos, Raja
Software Stack Intel oneAPI, Cray Shasta
Platform Cray Shasta