Cuda

Cuda is a GPU programming language developed by NVIDIA. Currently Shamu contains two (4) very high end GPU nodes consisting of eight (8) Telsa K80 GPU computational units and eight (8) V100s. The following instructions will show you how to access these nodes, where to find the sample code from the official SDK package and compiling examples. The GPU nodes should only be used for GPU programming, compilation and execution.

Access to the GPU queue

In order to access the GPU nodes you will need to request access by emailing your information (My UTSA ID) to rcsg@utsa.edu. If you already have an account on Shamu we will still need your sponsor or PIs permission to gain access to the GPU nodes.

-- AdminUser - 17 Jan 2017