New systems at ARL and ERDC DSRCs will provide an additional 10 petaFLOPS of computational capability.
The Department of Defense (DoD) High Performance Computing Modernization Program (HPCMP) completed its fiscal year 2016 investment in supercomputing capability supporting the DoD Science and Technology (S&T), Test and Evaluation (T&E), and Acquisition Engineering communities. The acquisition consists of three supercomputing systems with corresponding hardware and software maintenance services. At almost 10 petaFLOPs, this procurement will increase the DoD HPCMP's aggregate supercomputing capability to 31.1 petaFLOPs.
The new supercomputers will be installed at the Army Research Lab (ARL) and Engineer Research and Development Center (ERDC) DoD Supercomputing Resource Centers (DSRCs), and will serve users from all of the services and agencies of the Department:
- The Army Research Laboratory DSRC in Aberdeen, Maryland, will receive two SGI ICE X systems containing Intel Xeon "Broadwell" processors and NVIDIA Tesla K40 General-Purpose Graphics Processing Units (GPGPUs). These systems will consist of 33,088 and 73,920 "Broadwell" compute cores respectively, 32 GPGPUs each, 200 and 252 terabytes of memory respectively, 3.5 and 12 petabytes of disk storage respectively, and will provide 1.2 and 2.6 petaFLOPS of peak computing capability each.
- The US Army Corps of Engineers Engineer Research and Development Center DSRC in Vicksburg, Mississippi, will receive a Cray XC40 system containing Intel Xeon "Broadwell" processors, Intel Knights Landing (KNL) Many Integrated Core processors, and NVIDIA Tesla K40 GPGPUs. The system will consist of 126,192 "Broadwell" compute cores, 540 KNL nodes (64 cores each, 34,560 cores total), 32 GPGPUs, 437 terabytes of memory, 16 petabytes of disk storage, and will provide 6.05 petaFLOPS of peak computing capability.
The systems are expected to enter production service in the first half of calendar year 2017.