Upcoming Systems

Implementation Schedule
SiteEstimated DateDescription
AFRL DSRC Late Q3 FY 2021 Warhawk, a 5.1 peak petaFLOPS system, will contain 139,776 AMD EPYC (Rome) processor cores, 104 NVIDIA Volta V100 GPGPUs, a Cray Slingshot network (200 gb/s speed), 561 Terabytes of memory, and 24.84 petabytes of usable storage.
ARL DSRC Mid FY 2021 Jean will contain 57,696 Intel XEON "Cascade Lake Advanced Performance" compute cores and 280 NVIDIA Ampere A100 General-Purpose Graphics Processing Units (GPGPUs), interconnected with a 200 Gigabit per second InfiniBand network and supported by 323 terabytes of memory, and 12.5 petabytes of usable non-volatile memory express (NVMe)-based solid state storage.
ARL DSRC Mid FY 2021 Kay will contain 48,480 Intel XEON "Cascade Lake Advanced Performance" compute cores and 76 NVIDIA Ampere A100 General-Purpose Graphics Processing Units (GPGPUs), interconnected with a 200 Gigabit per second InfiniBand network and supported by 240 terabytes of memory, and 10 petabytes of usable non-volatile memory express (NVMe)-based solid state storage.
ERDC DSRC Mid CY 2021 Wheat will contain 890 Intel 9242 Cascade Lake AP compute nodes with 96 cores per node (85,440 cores). 770 nodes will be standard compute nodes, with 384 GBytes of memory per node; 16 will be visualization nodes, with 768 GBytes of memory and one nVidia A100 (Ampere) GPU each; 96 will be AI/ML nodes, with 384 GBytes of memory and one nVidia A100 (Ampere) GPU each; 8 will be large-memory nodes, with 2 TBytes of memory each. The /home file system will have 720 TBytes of usable NVMe storage, and the /app file system will have 360 TBytes.
Navy DSRC Early FY 2022 A Penguin Computing Open Compute Platform system with 176,128 total compute cores comprised of future generation AMD EPYC processors, and 144 NVIDIA Ampere A100 GPUs interconnected by a 200 gigabit per second Mellanox HDR-200 InfiniBand network, and supported by 26 PBytes of usable Data Direct Networks storage, including over 4 PBytes of NVMe-based solid-state storage, and 382 TBytes of system memory.
AFRL DSRC Early FY 2022 A Penguin Computing Open Compute Platform system with 189,440 total compute cores comprised of future generation AMD EPYC processors, and 152 NVIDIA Ampere A100 GPUs interconnected by a 200 gigabit per second Mellanox HDR-200 InfiniBand network, and supported by 26 PBytes of usable Data Direct Networks storage and 455 TBytes of system memory.