The NSUF provides scientific computing capabilities to support efforts in advanced modeling and simulation. These resources support a wide range of research activities, including performance of materials in harsh environments (including the effects of irradiation and high temperatures), performance of existing light water and advanced nuclear reactors, and multiscale multiphysics analysis of nuclear fuel performance.
Idaho National Laboratory
Currently, Idaho National Laboratory has two HPC resources:
Falcon: a SGI ICE-X distributed memory system with 34,992 cores, 121 TB of memory and a LINPACK rating of 1 Petaflop/s. Falcon’s network is a seven-dimensional enhanced hypercube utilizing FDR InfiniBand. Individual compute nodes contain dual Xeon E5-2695 v4 processors with 18 cores each and 128GB of memory. Falcon came online in Fall 2014 and ranked #97 on the November 2014 TOP500 list.
Lehmi: a Dell 6420-based system with 20,160 cores, 94 TB of memory and a LINPACK rating of 1 Petaflop/s. Lemhi’s network is an omnipath fat tree. Individual compute nodes contain dual Xeon Gold 6148 processors with 20 cores each and 192GB of memory. Lemhi came online in Fall 2018 and ranked #427 on the November 2018 TOP500 list.
Sawtooth: an HPE SGI 8600-based system with 99,792 cores, 403 TB of memory and a LINPACK rating of 5.6 Petaflop/s. Sawtooth's network is a nine-dimensional enhanced hypercube utilizing EDR and HDR InfiniBand. Individual compute nodes contain dual Xeon Platinum 8268 processors with 24 cores each. The majority of compute nodes contain 196 GB of memory while twenty-seven contain 384 GB of memory coupled with four NVIDIA V100 GPUs with 32 GB of on-GPU memory each. Sawtooth came online in late 2019 and ranked #37 on the November 2019 TOP500 list.
Computing resources are available to users who are proposing rapid-turnaround experiments or through the Consolidated Innovative Nuclear Research Funding Opportunity Announcement (CINR FOA). Rapid Turnaround Experiment (RTE) requests are accepted for up to 1 million core hours per proposal. Please refer to INL’s Nuclear Computational Resource Center for more information.
Technical Point of Contact: Eric Whiting (Eric.Whiting@inl.gov or 208-526-1433)