On Gordon (SDSC), what are the "bigflash" compute nodes?
On the Extreme Science and Engineering Discovery Environment (XSEDE), Gordon (SDSC) has four compute nodes available with the "bigflash" property. Each of these nodes has access to a separate 4.4 TB solid-state drive (SSD) file system. The "bigflash" compute nodes are intended for jobs needing larger scratch space than the 280 GB available to Gordon's native (non-vSMP) nodes.
To run a "bigflash" job on Gordon, change your PBS node request line:
#PBS -l nodes=number_of_nodes:ppn=16:native:bigflash
If your job needs more than one "bigflash" node, add this line:
#PBS -v Catalina_maxhops=2
The latency to the SSDs is several orders of magnitude lower than that for spinning disk, which makes them ideal for user-level check pointing, staging of data files that are accessed multiple times through the course of a job, and applications that need fast random I/O to large scratch files.
For more about XSEDE compute, advanced visualization, storage, and special purpose systems, see the Resources Overview, Systems Monitor, and User Guides. For scheduled maintenance windows, outages, and other announcements related to XSEDE digital services, see User News.
This is document bcua in the Knowledge Base.
Last modified on 2015-08-20 00:00:00.
- Fill out this form to submit your issue to the UITS Support Center.
- Please note that you must be affiliated with Indiana University to receive support.
- All fields are required.