Skip to main content

iForge

iForge is a high-performance computing cluster designed specifically for NCSA’s industry partners, featuring distinct hardware platforms designed for differing computational needs.

Researchers in industry can now solve their most complex modeling and simulation challenges faster. The powerful iForge cluster provides our partners in industry with the compute resources and user support they need to coax efficient performance and increased productivity out of even the most complex industrial applications. iForge’s GPU queue builds on this mission by offering fast new technology that’s already solving bigger problems in less time. With the latest NVIDIA V100 GPUs and NVlink interconnect, our partner engineers and scientists can get more from their machine learning and engineering applications. Initial benchmarks are already showing the benefits of running applications on the iForge GPU queue, which currently offers two 40-core, 192 GB Skylake nodes, each equipped with four NVIDIA Tesla V100 32 GB GPU cards (for a total of eight GPUs).

Architecture

BIG-MEM

  • Haswell
  • 24 nodes
  • 24 cores each
  • 256 GB RAM
  • $0.1308 per core-hour

Sklylake

  • Skylake
  • 44 nodes
  • 40 cores each
  • 192 GB RAM
  • $0.0947 per core-hour

GPU

  • Skylake
  • 2 nodes
  • 40 cores, 4 NVIDIA V100 GPUs in each
  • 192 GB RAM
  • $0.1931 per core-hour

COMPUTE RESOURCES ARE AVAILABLE TO A BROAD SLATE OF RESEARCHERS WORLDWIDE

Back to top