cc_staff
220
edits
Graham: Difference between revisions
Jump to navigation
Jump to search
no edit summary
m (Updated memory available on "new" nodes.) |
No edit summary |
||
| Line 172: | Line 172: | ||
<!--T:65--> | <!--T:65--> | ||
The two newest Volta nodes have 40 cores so the number of cores requested per GPU should be adjusted upwards accordingly, i.e. you can use 5 CPU cores per GPU. They also have NVLINK, which can provide huge benefits for situations where memory bandwidth between GPUs is the bottleneck. To use one of these NVLINK nodes, it should be requested directly, | The two newest Volta nodes have 40 cores so the number of cores requested per GPU should be adjusted upwards accordingly, i.e. you can use 5 CPU cores per GPU. They also have NVLINK, which can provide huge benefits for situations where memory bandwidth between GPUs is the bottleneck. To use one of these NVLINK nodes, it should be requested directly, add the switch "--constraint=cascade,v100" to the job submission script. | ||
<!--T:53--> | <!--T:53--> | ||