Category:Cluster: Difference between revisions

Latest revision as of 20:20, 17 July 2026

Current Usage Narrative

As of July 17, 2026:

Cluster 2 is your default choice for local compute. Most things work here.
Cluster 3 aka Wynton is your default for "at scale" calculations. Most of ZINC and most software is here. Cluster 3 will be going away in 2027. We'

're not sure exactly when.

Cluster 7 should be your default choice for local compute, except it still lacks some features and software. This is the cluster that will grow most between now and 2030 assuming we survive our leaders.
Cluster 8 will be your default choice for "at scale" calculations by 2027. For now, you can get access, and there is spare capacity.
Cluster 0 still exists, but is "legacy" and is being wound up.
Cluster 6 aka Keiser still exists, but is "legacy" and is being wound up.
Cluster 1 has not existed since 2014. ixnay onyay ethay uestionsqay.

Set up your own cluster

see: So you want to set up a lab

New group members

please use Cluster 7. Most group members as of Feb 2026 are using Cluster 2. A few are using Cluster 7, just ask us. Cluster 8 (coreHPC) is available. exceptions.

- the cpu/gpu and disk spaces are guesses and no effort has been made to make them accurate. If you know the correct numbers, please update and indicate this **

Cluster ID	Description	How to get started	cpu/gpu	Disk space	Notes
Cluster 0	Legacy UCSF cluster	Closed	300/0	20 TB	Legacy support. Retire asap.
Cluster 1	Toronto Cluster	Closed	448/0	40 TB	Defunct
Cluster 2	New UCSF cluster	Closed	800/52	6.5 PB	No new growth. Merge into Cluster 7
Cluster 3	UCSF Shared Cluster aka Wynton	Closed	18,000/250	6 PB	Wynton is end of life. Merging with cluster 8.
Cluster 5	Amazon Cloud	ask John	unlimited	unlimited	Commercial. Ask us. Watch out for bandwidth costs.
Cluster 6	Mike Keiser cluster	Closed	100/50	1 PB	End of life. being merged into cluster 7
Cluster 7	New Cluster 7	JJ/Khanh	100/8	1 PB	Future main private cluster for our group
Cluster 8	FAC/coreHPC	UCSF IT	16,000/350	1 PB	UCSF replacement for Wynton
Cluster 9	Offsite experimental cluster	John	64/4	1 PB	off-site cluster

Access our clusters from secure machines only. If unsure, ask the sysadmins.

Pages in category "Cluster"

The following 17 pages are in this category, out of 17 total.

A

About our cluster

C

H

Hardware

O

OGS/GE

S

T

U

Using SGE cluster

@@ Line 1: / Line 1: @@
-We currently have three computer clusters, but aim to reduce this to two or even one.
+== Current Usage Narrative ==
-* [[Cluster 0]] also known as the legacy cluster is in San Francisco at UCSF.  (was UCSF Cluster)
+As of July 17, 2026:
-* [[Cluster 1]] is in the pharmacy building at U of T.  (was UofT Cluster)
+* Cluster 2 is your default choice for local compute. Most things work here.
-* [[Cluster 2]] is in Byers Hall 101 and Genentech Hall 108 at UCSF aka "the new cluster".
+* Cluster 3 aka Wynton is your default for "at scale" calculations. Most of ZINC and most software is here.  Cluster 3 will be going away in 2027. We'
+'re not sure exactly when.
+* Cluster 7 should be your default choice for local compute, except it still lacks some features and software. This is the cluster that will grow most between now and 2030 assuming we survive our leaders.
+* Cluster 8 will be your default choice for "at scale" calculations by 2027. For now, you can get access, and there is spare capacity.
+* Cluster 0 still exists, but is "legacy" and is being wound up.
+* Cluster 6 aka Keiser still exists, but is "legacy" and is being wound up.
+* Cluster 1 has not existed since 2014. ixnay onyay ethay uestionsqay.
-Access clusters via sgehead if on campus, [[Portal system | portals]] if off campus.
+== Set up your own cluster ==
+see: [[So you want to set up a lab]]
-We also have access to the QB3 shared cluster:
+== New group members ==
+please use [[Cluster 7]].
+Most group members as of Feb 2026 are using [[Cluster 2]].
+A few are using [[Cluster 7]], just ask us. Cluster 8 (coreHPC) is available.
+[[cluster exceptions|exceptions]].
-* [https://salilab.org/qb3cluster QB3 Shared Cluster] aka [[Cluster 3]].
+** the cpu/gpu and disk spaces are guesses and no effort has been made to make them accurate. If you know the correct numbers, please update and indicate this **
-We can also use commercial cloud clusters
+{| class="wikitable"
-*  Amazon cloud EC2, which we affectionately refer to as [[Cluster 5]]
+|-
-* Google cloud, which we call [[Cluster 6]].
+! Cluster ID !! Description !! How to get started || cpu/gpu || Disk space || Notes
+|-
+| [[Cluster 0]] ||  Legacy UCSF cluster || Closed || 300/0 || 20 TB ||  Legacy support. Retire asap.
+|-
+| [[Cluster 1]] || Toronto Cluster || Closed||  448/0 ||40 TB || Defunct
+|-
+| [[Cluster 2]] ||  New UCSF cluster || Closed || 800/52 || 6.5 PB || No new growth. Merge into Cluster 7
+|-
+| [[Cluster 3]]  || [https://salilab.org/qb3cluster UCSF Shared Cluster] aka Wynton ||  Closed || 18,000/250 || 6 PB ||  Wynton is end of life. Merging with cluster 8.
+|-
+| [[Cluster 5]]  || [https://aws.amazon.com/ Amazon Cloud] ||  ask John  || unlimited || unlimited ||  Commercial. Ask us. Watch out for bandwidth costs.
+|-
+| [[Cluster 6]]  || Mike Keiser cluster ||  Closed || 100/50 || 1 PB ||  End of life. being merged into cluster 7
+|-
+| [[Cluster 7]]  || New Cluster 7 ||  JJ/Khanh || 100/8 || 1 PB ||  Future main private cluster for our group
+|-
+| [[Cluster 8]]  || FAC/coreHPC ||  UCSF IT || 16,000/350 || 1 PB ||  UCSF replacement for Wynton
+|-
+| [[Cluster 9]]  || Offsite experimental cluster ||  John || 64/4 || 1 PB ||  off-site cluster
+|}
-Need advice about which cluster to use? Ask the [[sysadmin]].
+Access our clusters from [[Lab Security Policy | secure machines]] only.  If unsure, ask the [[sysadmin]]s.
 [[Category:Internal]]
-[[Category:Cluster]]
+[[Category:Topic]]