High-performance Computing
Clusters
The Genome Center Bioinformatics computing clusters consist of an ever evolving group of clusters collaboratively administered by the Bioinformatics Core Group housed at the UC Davis Genome Center. The clusters are utilized as a collective resource for serial and parallel applications that would be too computationally demanding for smaller research groups to implement. Where one researcher could purchase a small cluster in a grant and hire a system administrator to set it up, it is much more efficient to add computing power to existing infrastructure. This is where the Bioinformatics clusters come into play.
A few examples of large scale problems the clusters are regularly used for include:
- Large Scale Sequence Alignment
- Hidden Markov Model Development and Searches
- 3D Molecular modeling
- Mass Spec models
- Phylogenetic inference
Cluster etiquette and usage resources
A presentation about Genome Center cluster computing concepts and usage: Cluster_presentation.pdf
A simple tutorial on use of the Sun Grid Engine (the job scheduler in ROCKS) to run jobs on the clusters (from Danish Technical University)
Ganglia cluster toolkit (GUI for monitoring job load, etc) on Shiraz (requires cluster account login)
UCSC genomewiki about cluster jobs (for parasol, not ROCKS)
- Shiraz
- 111 nodes, dual socket dual core Opteron running Rocks 4.1
- 2 Opteron 270 (2.0 GHz) processors per node
- 37 nodes w/ 8GB Ram, the rest have 4GB per node
- 6.4 TB (raw) backed up redundant storage
- Apple
- 37 nodes, Dual G4 running OSX 10.3
- Dual 1 GHz processors, 2G RAM per node
- 1.8 TB (raw) backed up redundant storage
- Genbeo
- 24 nodes, dual Opteron running Rocks 4.1
- Dual Opteron 248 (2.2 GHz) processors, 4GB RAM per node
- 3.2 TB (raw) backed up redundant storage
- Voignier?? In Development
- 7 nodes, Itanium 1/2?? running Rocks 4.2 Beta
- Dual Itanium (x.x GHz) processors, 4GB RAM per node
- 3.2?? TB (raw) backed up redundant storage