Hardware

Trixie is a GPU cluster consisting of 36 nodes, each with NVIDIA V100 GPU a fast, Infiniband Interconnect, and a large 1 PB global filesysetm

Operating system

Runs CentOS 7

Job scheduler

https://slurm.schedmd.com (for example run scripts on Trixie see Running-jobs)

Headnode (2 available)

• processor_type = Intel Xeon Gold 6130 CPU clocked at 2.1GHZ 16 cores / CPU
• processors_per_node = 2
• RAM = 96 GB memory

Node Profile

• processor_type = Intel Xeon Gold 6130 CPU clocked at 2.1GHZ 16 cores / CPU
• processors_per_node = 2
• RAM = 192 GB memory
• GPU_type = NVidia V100 GPUs with 32 GB RAM / GPU
• GPU_details = https://www.nvidia.com/en-us/data-center/tesla-v100/Nodes
• GPU_per_node = 4
• local scratch size =

Filesystem

• 1 PB of GPFS storage total, with the following mount points and quotas

• /home/username User space. Not backed up. Soft quota 500 GB.

• /project/PROGRAM/project_id Shared space for sharing files between users within a project. Each project has a default quota of 1 TB. If additional space is needed projects can make an official request. Space is not backed up.

• /scratch/username User controlled space for running jobs. Not backed up. Subject to Purge policy