For the complete documentation index, see llms.txt. This page is also available as Markdown.
maxwell
Maxwell is a SuperMicro AS-4124-GS-TNR chassis with 256 HT Cores (2 sockets with 64 cores/socket and 2 threads/core) based on the AMD EPYC 7763 64-Core Processor, and 1 TB main memory. It has 3 TB nvme scratch storage.
This system has four Nvidia A100-PCIE-40GB GPUs and four AMD Instinct MI100 GPUs.
The ROCM development environment does not require a module load; it is installed in system space.
For Nvidia:
moduleloadnvhpc
to set up all of the environment needed.
A very light test program is available via git at https://github.com/jungwonkim/amd-toy. This is a good way to ensure your environment is set up correctly.
All tests should return err[0]. If they do not, then it is likely that you do not have render group permissions
To check, run the groups command (on maxwell) and see if you are in the render group.
