1. Stella
  2. Hardware
Getting Access

To request access to this resource, please email hpcteam@temple.edu with a brief explanation of your intended use. If you are a student or postdoctoral researcher, be sure to include the name of your supervisor in your request.

This page is under development!

Hardware

Microway Numbersmasher NVIDIA DGX-1V
Compute nodes
GPU nodes
CPU 2x AMD EPYC 7543, 64-core, 2.8GHz Base 2x AMD EPYC 7343, 32-core, 3.2GHz Base
RAM 512 GBz, RDIMM DDR4 256 GB, RDIMM DDR4
GPU 3x NVIDIA A100 Tensor Core per node
  • 9.7 teraflops, FP64
  • 19.5 teraflops, FP64 Tensor Core
  • 624 teraflops, FP16 Tensor Core
  • 80GB memory per GPU
  • 3x 8192 NVIDIA CUDA Cores
  • 3x 512 NVIDIA Tensor Cores | Gen 3
GPU Interconnect NVlink 2
  • NVLink connects
  • 50GB/s bi-directional bandwidth per link