TACC User Guides
Last update: August 10, 2020
Largest open-science cluster in the U.S.
8008 Intel Cascade Lake processors
Bare-metal & VM cloud research
Multi-architecture: Haswell, ARM, FPGA, GPUs
System components at UChicago and TACC
High Performance Storage and Data Management
6PB replicated storage in Austin and Arlington
Direct access on Stampede2, Lonestar5, Stallion
Self-service cloud and virtualization
System components at Indiana U and TACC
HPC, Remote Visualization
Cray XC40 Cluster with Intel Xeon 1.2 PF, 30K cores; Haswell-based Cray XC40
HPC, Deep Learning
Longhorn supports GPU-accelerated jobs, with single-GPU and multi-GPU nodes.
Machine Learning/Deep Learning
30 GPU nodes, (GTX, V100 and P100 types)
Mass Archival Storage
The TACC mass storage server provides 160PB of both short term and long term high performance archival storage.
Dell Visualization Cluster; 16x5, 328 megapixel tiled display
Wrangler Decommissioned September, 2019
Data Intensive Computing
120 Haswell nodes; EMC DSSD flash high-speed global store