Wombat is a single rack cluster from HPE based on the 64-bit ARM architecture instead of traditional x86-based architecture. This system is available to support computer science research projects aimed at exploring the ARM architecture. Appropriate research topics could include, but are not limited to, functionality testing of research software and prototypes, power-efficiency evaluations, and efforts to improve the system software, runtime support and the programming environment for ARM architecture based systems.
The Wombat cluster is a heterogeneous cluster with 24 compute nodes. 16 compute nodes are based on the Fujitsu A64fx processor, 4 compute nodes have Marvell ThunderX2 processors plus Nvidia Volta GPU’s and the last 4 nodes have ThunderX2 processors but no GPU’s. The A64fx processor has 48 cores and 32GB of on-chip HBM2 memory. Each of the 16 nodes has one processor socket and a 512 GB NVME device for node-local storage. The A64fx-equipped nodes do not have additional DDR memory. The ThunderX2 processors have 32 cores. Each of the 8 nodes has 2 processor sockets (64 cores total), 256 GB of DDR4 RAM and a 480 GB SSD for node-local storage. The 4 nodes equipped with GPU’s have 2 GPU’s each. (8 GPU’s total in the cluster.) The GPU’s are Nvidia Volta GV100 accelerators with 32 GB HBM2 memory. All nodes are connected with EDR InfiniBand (~100Gbit/s).
To access Wombat, users must apply for an OLCF account. (See the link below.) Once an account has been created, users can log in to Wombat by ssh’ing to wombat-login1.ccs.ornl.gov.
When filling out the “Project Information” section of the account application, enter “Ross Miller” as the PI, CSC266 as the project name and select “Requesting New Membership” under project membership status. Please choose either bash or csh for your preferred shell. Fill out the rest of the form appropriately and click the “Submit” button. (Don’t forget to check the “Accurate and Complete” checkbox at the bottom before submitting.)
Account applications are available here: Account Application.
Please note that all published work related to Wombat should include the following acknowledgement statement:
This research used the resources of the Oak Ridge Leadership Computing Facility, located in the National Center for Computational Sciences at the Oak Ridge National Laboratory, which is managed by UT Battelle, LLC for the U.S. Department of Energy, under the contract No. DEAC05-00OR22725.