Difference between revisions of "RRA Hardware"

(One intermediate revision by the same user not shown)
Line 1: Line 1:
= RRA Hardware =
= RRA Hardware =


Advanced computing resources at the University of South Florida are administered by Research Computing (RC). RC hosts the RRA (Restricted Research Access) cluster computer which currently consists of approximately 77 nodes with over 1,600 processor cores running Red Hat Enterprise Linux v7. The cluster is built on the [https://wiki.rc.usf.edu/index.php/What_is_the_condominium_model condominium model] with 16TB of memory shared across the nodes in various configurations. Additionally, there are 120 GPUs available. The nodes utilize either 100GB Omnipath or 100GB Infiniband for a computational interconnect. There is a 350TB encrypted BeeGFS parallel file system used to support high speed and I/O intensive computations, as well as for long-term storage. RC also provides and supports more than 120 scientific software packages for use on a variety of platforms. Remote system and file access is available from essentially anywhere. RC staff members are available to facilitate use of the cluster, as well as provide direct assistance with research projects that require high-performance computing or advanced visualization and analysis of data. User education and training sessions are also provided several times per semester.
Advanced computing resources at the University of South Florida are administered by Research Computing (RC). RC hosts the RRA (Restricted Research Access) cluster computer which currently consists of approximately 77 nodes with over 1,600 processor cores running Red Hat Enterprise Linux v7. The cluster is built on the [https://wiki.rc.usf.edu/index.php/What_is_the_condominium_model condominium model] with 16TB of memory shared across the nodes in various configurations. Additionally, there are 120 GPUs available. The nodes utilize either 100GB Omnipath or 100GB Infiniband for a computational interconnect. There is a 350TB encrypted BeeGFS parallel file system used to support high speed and I/O intensive computations, as well as for long-term storage. RC also provides and supports more than 120 scientific software packages for use on a variety of platforms. Remote system and file access is available from essentially anywhere. RC staff members are available to facilitate use of the cluster, as well as provide direct assistance with research projects that require high-performance computing or advanced visualization and analysis of data. '''User education and training sessions are also provided upon request.'''


* The above statement as well as any information below can be used as part of the facilities description for grant proposals, provided that it is acknowledged that these resources are administered by Research Computing at the University of South Florida.
* The above statement as well as any information below can be used as part of the facilities description for grant proposals, provided that it is acknowledged that these resources are administered by Research Computing at the University of South Florida.
Line 70: Line 70:


* [[SLURM_Partitions|CIRCE Partition Layout]]
* [[SLURM_Partitions|CIRCE Partition Layout]]
* [[RRA Layout|RRA File System Layout]]
* [[RRA layout|RRA File System Layout]]
* [[RRA Data Management|RRA Data Management]]
* [[RRA Data Management|RRA Data Management]]

Revision as of 14:48, 10 May 2021

RRA Hardware

Advanced computing resources at the University of South Florida are administered by Research Computing (RC). RC hosts the RRA (Restricted Research Access) cluster computer which currently consists of approximately 77 nodes with over 1,600 processor cores running Red Hat Enterprise Linux v7. The cluster is built on the condominium model with 16TB of memory shared across the nodes in various configurations. Additionally, there are 120 GPUs available. The nodes utilize either 100GB Omnipath or 100GB Infiniband for a computational interconnect. There is a 350TB encrypted BeeGFS parallel file system used to support high speed and I/O intensive computations, as well as for long-term storage. RC also provides and supports more than 120 scientific software packages for use on a variety of platforms. Remote system and file access is available from essentially anywhere. RC staff members are available to facilitate use of the cluster, as well as provide direct assistance with research projects that require high-performance computing or advanced visualization and analysis of data. User education and training sessions are also provided upon request.

  • The above statement as well as any information below can be used as part of the facilities description for grant proposals, provided that it is acknowledged that these resources are administered by Research Computing at the University of South Florida.

Server Hardware

Partition Nodes Core Count Processors Cores per Node Memory per Node GPUs per Node Interconnect
rra 55 1100 2x Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz 20 192GB 1x GTX 1080 Ti 100GB Omnipath
rra 19 456 2x Intel(R) Xeon(R) Silver 4214 CPU @ 2.20GHz 24 192GB 3x RTX 2080 Ti 100GB Infiniband
rra_con2020 2 64 2x Intel(R) Xeon(R) Silver 4214 CPU @ 2.20GHz 32 192GB 4x Tesla V100 100GB Infiniband
rra_con2020 1 48 4x Intel(R) Xeon(R) Gold 6246 CPU @ 3.30GHz 48 1.5TB none 100GB Infiniband
Totals 77 1668 16092 GB 120



Additional Information for Cluster Hardware and Usage