How to Cite #

Grant Applications #

Below are a few examples on how to give details on the Wynton HPC environment in a grant application. The node, CPU, and GPU core counts are updated on a regular basis.

Example 1 #

Wynton HPC is a distributed high-performance computing cluster with nodes in many different data centers on the UCSF campus. All Wynton HPC sites are connected by multiple 40 Gbps network connections. Wynton HPC currently contains 502 nodes with over 17600 CPU cores. RAM in the nodes ranges from 48 to 1512 GiB with an average RAM-to-core ratio of over 10 GiB. There are also 61 nodes containing a total of 235 GPUs. Storage is provided by a parallel filesystem providing 400 TB of mirrored home space and 500 TB of global scratch space.

Wynton is shared by many research groups at UCSF. Load balancing on the cluster is achieved through SGE (originally Sun Grid Engine). The computing power of the cluster guaranteed to any given research group is directly proportional to the funds that the group contributed to the cluster. The balance of the computing power not used by the contributors is available to other registered groups (including other contributors) on a first-come, first-served basis.

Example 2 #

Wynton HPC is a computational research cluster shared by investigators at UCSF. The cluster is run as a co-op, with access for all and priority given to those who have contributed funds or hardware. Support is also provided by UCSF Research IT. The cluster currently consists of 502 nodes with 17600 cores. Each node has at least 48 GiB of RAM and total home storage is 770 TiB.