Semiconductor & Computer Engineering
Aurora is a planned state-of-the-art exascale supercomputer designed by Intel/Cray for the U.S. Department of Energy's (DOE) Argonne Leadership Computing Facility (ALCF). The system is expected to become the first supercomputer in the United States to break the exaFLOPS barrier.
Contents
History
Originally announced in April 2015, Aurora was planned to be delivered in 2018 and have a peak performance of 180 petaFLOPS. The system was expected to be the world's most powerful system at the time. The system was intended to be built by Cray based on Intel's 3rd generation Xeon Phi (Knights Hill microarchitecture). In November 2017 Intel announced that Aurora has been shifted to 2021 and will be scaled up to 1 exaFLOPS. The system will likely become the first supercomputer in the United States to break the exaFLOPS barrier. As part of the announcement Knights Hill was cancelled and instead be replaced by a "new platform and new microarchitecture specifically designed for exascale".
Original Specifications
| Original Specs | |
|---|---|
| Computing Power | 180 PFLOPS |
| Compute Nodes | >50,000 |
| Processor | 3rd Generation Intel Xeon Phi (Knights Hill) |
| Memory | >7 PB DRAM and persistent memory |
| Fabric | 2nd Generation Intel Omni-Path Architecture with silicon photonics |
| File System | Intel® Lustre* File System |
| File System Capacity | >150 Petabytes |
| File System Throughput | >1 Terabyte/s |
| Peak Power | 13 MW |
| FLOP/s Per Watt | >13 GigaFLOP/s per watt |
| Facility Area | ~3,000 sq. ft. |
Documents
Original System
- Ushering in a New Era, Argonne National Laboratory’s Aurora System, April 2015
- Aurora Fact Sheet
External links
References
- Introducing Aurora, April 9, 2015
- U.S. Department of Energy Selects Intel to Deliver Nation’s Most Powerful Supercomputer at Argonne National Laboratory, April 9, 2015
- Unleashing High-Performance Computing Today and Tomorrow, November 13, 2017