User Tools

Site Tools


cuda

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision
Previous revision
cuda [2015/07/19 10:31]
volker created
cuda [2015/07/19 10:35] (current)
volker
Line 8: Line 8:
  
 where XX is the Compute Capability of the cards the code will run on. For cards available at the ICS, use the following. where XX is the Compute Capability of the cards the code will run on. For cards available at the ICS, use the following.
- 
  
 ^Slurm Partition^Card^Compute Capability^ ^Slurm Partition^Card^Compute Capability^
-|tasna|Geforce ​GTX 590|20| +|tasna|GeForce ​GTX 590|20| 
-|ceres|Geforce ​GTX 680|30| +|ceres|GeForce ​GTX 680|30| 
-|vesta|Tesla ​K80|37| +|vesta|Tesla ​K80X|37|
  
 +:!: **You should compile on nodes of the partition you are running on. To do so, launch an interactive job as described on the [[slurm|SLURM Scheduler]] page.** :!:
  
 +Note that the GTX 680 cards have gimped double precision performance compared to the GTX 590. Do not be surprised if your code runs slower than you would expect.
cuda.1437294665.txt.gz · Last modified: 2015/07/19 10:31 by volker