Pegasus and Triton Queues
LSF queues are organized using limits like job size, job length, job purpose, and project. In general, users run jobs on the clusters with equal resource shares. A user’s current or recent resource usage lowers the priority applied when LSF assigns resources for their new jobs.
The normal queue is available for both serial and parallel jobs.
The bigmem queue, on Pegasus, is available for jobs requiring more than 50 GB of memory per node. Submitting jobs to this queue requires project membership. Do not submit jobs that can run on the normal queues to the bigmem queue.
Warning
Jobs using less than 1.5G of memory on the bigmem queue are in violation of acceptable use policies and the IDSC account responsible for those jobs may be suspended.
Pegasus Queues
Details:
Queue Name |
Cores per node |
Memory |
Wall time default / max |
Description |
|---|---|---|---|---|
normal |
16 |
up to 50 GB per node |
1 day / 7 days |
jobs intended for workloads that run entirely on a single node. |
parallel |
16 per node |
up to 50 GB per node |
1 day / 7 days |
jobs intended for MPI or other multi-node parallel workloads. |
bigmem |
16 |
up to 200 GB per node |
4 hours / 5 days |
jobs requiring nodes with more than 200 GB of memory, max. of 32 cores allowed per job. |
debug |
16 |
up to 32 GB |
30 mins / 30 mins |
job debugging, max. of 64 cores allowed per job. |
interactive |
16 (15 for jobs) |
up to 250 GB per node |
6 hours / 1 day |
interactive jobs |
gpu_h100 |
2 GPUs |
96 GB max V-RAM/node |
1 day / 7 days |
gpu h100 |
gpu_titan |
2 GPU |
12 GB max V-RAM |
1 day / 7 days |
gpu titan v100 |
Triton Queues
Details:
Queue Name |
Cores per node |
Memory |
GPU |
Description |
|---|---|---|---|---|
normal |
40 |
up to 280 GB per node |
2 V100s with 16GB VRAM and CUDA 12.4 |
parallel & serial jobs up to 16 cores, up to 50 GB memory |
interactive |
40 (15 for jobs) |
up to 280 GB per node |
2 V100s with 16GB VRAM and CUDA 12.4 |
interactive jobs |