Differences

This shows you the differences between two versions of the page.

--- partitions [2014/11/04 15:06]
root
+++ partitions [2025/03/20 13:02] (current)
root updated GPU offerings - kmd
@@ Line 3: / Line 3: @@
 **Partition** == **Queue**
-SLURM allows the definition of different partitions. Think of these as job queues.
+SLURM allows the definition of different partitions. A partition is a subset of all the nodes in the cluster. Partitions are not mutually exclusive, so a node can belong to multiple partitions. Every compute node is assigned to one or more partitions.
 Our cluster has 4 different partitions:
-  * **all**
+  * **standard**
-    * The default queue. Contains the 17 standard compute nodes (128GB, 16 cores each).
+    * The default queue. Contains the 31 standard compute nodes (at least 128GB RAM, at least 16 cores). No time limit.
   * **gpu**
-    * Contains the GPU machine only.
+    * Contains the GPU machines only. No time limit. Nodes node90 and node91 each have an NVIDIA GV100GL [Tesla V100 PCIe 16GB] GPU, and node94 has an NVIDIA Corporation GV100GL [Tesla V100 PCIe 32GB] GPU.
   * **bigmem**
-    * Contains the big memory machines only.
+    * Contains big memory machines only (512GB or 768GB, 32 cores). No time limit.
   * **short**
-    * Contains all nodes, including gpu and bigmem, but has a 2 hour time limit, and jobs in this queue have a lower priority than jobs in the other queues.
+    * Contains all nodes, including gpu and bigmem, but has a 5 hour time limit, and jobs in this queue have a lower priority than jobs in the other queues. The short queue allows users with quick-running jobs to make use of all the cluster nodes without holding up users who need the special purpose (bigmem or gpu) nodes too much.
-**All**, **gpu**, and **bigmem** have no time limit.
+**Standard**, **gpu**, and **bigmem** have no time limit.

BRC Cluster Workshop

User Tools

Site Tools

Differences

Page Tools