⚠️ 2021-11-01: The TIPCC cluster is retired. If you're looking for a modern cluster, please see our new C4 cluster, which is in production since January 11, 2021.

Cluster Specifications

Software environment

All nodes on the cluster runs Scyld ClusterWare 6.6 (Red Hat Enterprise Linux 6.6) by Scyld Software (subsididary of Penguin Computing). The job scheduler is Torque/PBS 4.2.9 which runs together with the Moab workload manager.

Hardware

The cluster can be accessed via SSH to the login node:

cclc01: cclc01.som.ucsf.edu

The login nodes can only be accessed from within the UCSF network.

Development Nodes

The cluster has development nodes for the purpose of validating scripts, prototyping pipelines, compiling software, and more. Development nodes can be accessed using SSH and only from the login nodes.

Node	# Cores	CPU	RAM	Local `/scratch`
n6	12	2.4 GHz	64 GiB	0.86 TiB

Compute Nodes

Node	# Cores	CPU	RAM	Local `/scratch/`	Priority
n8	12	2.3 GHz	64 GiB	0.86 TiB	(communal)
n11	12	2.3 GHz	64 GiB	1.8 TiB	(communal)
Total	24

The head node specs: 48 cores, 2.6 GHz CPU, 512 GiB RAM.
The local disk space is mounted as /scratch/, unique to each node and, and is much faster than any other disk on the cluster. Files on local /scratch and local /tmp older than 35 days are automatically removed. On interactive node n6, this limit is 14 days. All nodes have AMD processors. The 2 communal nodes with a total of 24 cores are available to all users. The non-communal nodes are only available to specific groups and their members.

Source: pbsnodes, cat /proc/cpuinfo and grep -F SRCFG /opt/moab/etc/moab.cfg.

Compute Nodes Migrated to C4

The following TIPCC compute nodes have been, or is in the process of being, migrated to the C4 cluster.