The BIO Cluster

Getting Started on the BioCluster:

Prof. Kevin R. Thornton has written the following helpful instructions on how to use the BioCluster on HPC:



The BioCluster is a condo cluster that is part of the HPC Cluster.   The BioCluster hardware is owned by the School of Biological Sciences at UCI. 

The BioCluster hardware consists of 11 BIO compute nodes with approximately ~350TB of Raided disk space in the following distribution:
Hardware
Owners
Five (5) 64-core nodes each with 512GB of memory
 School of Biological Sciences
Three (3) 100TB Raid Servers
 School of Biological Sciences
One (1) 64-core node with 512GB memory and 18TB data
 Prof. Kevin R. Thornton
One (1) 64-core node with 512GB memory and 18TB data
One 100TB Raid Server
 Prof. Seyed Ali Mortazavi
Three (3) 64-core nodes each with 256GB of memory
 Prof.  Robin M. Bush
One (1) 64-core node with 512GB memory
 Prof. Brandon Stuart GAUT


Click here for a picture of the BIO Compute & Data Rack.



Accessing the BIO Compute nodes:

The following Queues have been created on HPC for the BIO group with the following Q configurations:
Queue Name
  Group
Access
Nodes
  Suspend able Jobs?
bio
 bio
 5 64-core nodes ( 320 cores )
 No
sf
 sf
 3 64-core nodes ( 192 cores )
 No
krt
 krt
 1 64-core node ( 64 cores )
 No
sam
 samlab
 1 64-core node ( 64-cores )
 No
bsg
 bgaut
 1 64-core node ( 64-cores ) No
abio bio
sf
krt
samlab
bgaut
 Total possible cores = ( 704 cores )
 		YES


Which Queues can I access?

From your account, enter "groups" and see what groups you belong to.   For example:
$ groups
samlab  bio

The listing above says that I belong to the "samlab" and "bio" groups, so I can run on the "sam", "bio" and "abio" Queues. 

The all bio "abio" queue will use any nodes from bio, sf, krt, sam, or bsg queues only when nodes on those queues are not in use by their owners.    Jobs on the abio queue can be suspended while jobs on all queues listed above in black will never be suspended.

The reason for the abio queue is to allow all BIO users to share all BIO nodes when their owners are not using their nodes.  



BIO Data Storage Servers:
Mount Point
 Group
Access
 Size
Data server
Connection
/bio
 bio
sf
krt
samlab
bgaut
340TB
 Gluster File system
on 4 unique data servers
Infiniband
Network
/kevin
 krt 11TB
 compute-2-1
 Infiniband
Network
/samlab
 samlab 11TB
 compute-2-2
 Infiniband
Network



If you have questions and/or need help using the BioCluster, please send an email to the RCS Group,  Joseph Farran at jfarran@uci.edu, or to Harry Mangalam at hmangala@uci.edu