Tron is an ASC visualization and data analysis resource for users of the Sierra Tri-Lab computer. When not in use for data visualization or analsyis in support of a current ATCC cycle, the system will run compute work. Tron mounts the same file systems as Sierra.
There will be Visualization Dedicated Access Times (DAT) on short notice when needed. When possible, 1 hour of advance notice will be provided users.
Tron jobs are scheduled through SLURM and jobs are scheduled per node. Tron has two scheduling pools;
- pvis—60 nodes (math)
- pbatch—84 nodes (math)
Scheduling
Pvis nodes are targeted for Sierra visualization and data analysis.
Pbatch nodes are for postprocessing and other compute work.
Pools Max nodes/job Max runtime --------------------------------------------------- pvis 24 unlimited pbatch 24 24 hours ---------------------------------------------------
NOTE: Jobs in pvis WILL be killed if they do not meet this criteria;
1. run in a valid visualization bank provided to current ATCC users on Sierra
OR
2. run in standby
While there are currently no time limits for pvis, there may be limits established per user account or per bank. If needed, more restrictive limits will be established.
Hardware
- Intel Cascade Lake 2.9GHz cores
- Mellanox EDR Interconnect
- 32 cores/node
- 384GB memory/node
Documentation
- Linux Clusters Tutorial Part One | Linux Clusters Part Two
- Slurm Tutorial (formerly Slurm and Moab)
Contact
Please call or send email to the LC Hotline if you have questions. LC Hotline | phone: (925) 422-4531 | email: lc-hotline@llnl.gov
See Compilers page
Introduction to LC Resources
Linux Clusters Overview
Slurm Tutorial (previously Slurm and Moab)