TRACE is designed for converged HPC + AI + Data. Its custom topology is optimized for data-centric HPC, AI, and HPDA (High Performance Data Analytics).
TRACE nodes
Trace nodes provide extremely powerful general-purpose computing, pre- and post-processing, AI inferencing, and machine learning and data analytics.
trace nodes | ||
|---|---|---|
Number | 12 | |
CPU | 2 AMD EPYC 7713 CPUs 64 cores per CPU, 128 cores per node 2.0 GHz | |
RAM | 2TB memory (64GB RDIMM, 3200MT/s) | |
GPUs | ~150k CUDA cores |
Storage($LOCAL) | 3.2TB NVMe (1.6TB U.2 Gen4 RAID0) |
Storage
Network | Mellanox ConnectX-6 Dual Port 100Gb InfiniBand |