GPU-accelerated workloads

Longhorn is a TACC resource built in partnership with IBM to support GPU-accelerated workloads. The power of this system is in its multiple GPUs per node, and it is intended to support sophisticated workloads that require high GPU density. Longhorn supports double-precision machine learning and deep learning workloads that can be accelerated by GPU-powered frameworks, as well as general purpose GPU calculations.

 Longhorn User Guide

IBM Power System Specifications

Longhorn is composed of 108 IBM Power System AC922 nodes distributed across nine racks, plus an IBM Elastic Storage System as a standalone 10th rack for the Home and Scratch file systems. Four nodes are reserved as login nodes and management nodes, leaving 104 nodes for the compute system.

GPU Nodes — Longhorn hosts 96 V100 nodes, each with 4 GPUs per node.

GPU Large Memory Nodes — Longhorn hosts 8 large memory V100 nodes, each with 4 GPUs per node.