This shows you the differences between two versions of the page.
markdowntest [2022/02/08 23:10] root created |
markdowntest [2022/02/08 23:11] (current) root |
||
---|---|---|---|
Line 10: | Line 10: | ||
faster with the `+setcpuaffinity +isomalloc_sync` options. charmrun | faster with the `+setcpuaffinity +isomalloc_sync` options. charmrun | ||
`++ppn` should match PBS `ppn=`. | `++ppn` should match PBS `ppn=`. | ||
+ | |||
+ | *** | ||
module load namd/ | module load namd/ | ||
Line 21: | Line 23: | ||
charmrun ++remote-shell ssh ++ppn 16 `which namd2` \ | charmrun ++remote-shell ssh ++ppn 16 `which namd2` \ | ||
+p $NP +setcpuaffinity +isomalloc_sync apoa1.namd > | +p $NP +setcpuaffinity +isomalloc_sync apoa1.namd > | ||
+ | |||
+ | *** | ||
This is for single-node run using only the shared-memory program | This is for single-node run using only the shared-memory program | ||
Line 29: | Line 33: | ||
NP=$(wc -l < | NP=$(wc -l < | ||
namd2 +p $NP apoa1.namd +setcpuaffinity +isomalloc_sync > | namd2 +p $NP apoa1.namd +setcpuaffinity +isomalloc_sync > | ||
+ | |||
+ | *** | ||
##### Benchmarks | ##### Benchmarks | ||
Line 43: | Line 49: | ||
version didn't show any useful scaling for more than 2 nodes, and AMD | version didn't show any useful scaling for more than 2 nodes, and AMD | ||
not very useful scaling for more than 3 nodes. | not very useful scaling for more than 3 nodes. | ||
+ | |||
+ | *** | ||
Node Type ppn version | Node Type ppn version | ||
Line 53: | Line 61: | ||
32-core AMD 32 | 32-core AMD 32 | ||
32-core AMD 32 | 32-core AMD 32 | ||
+ | |||
+ | *** | ||
##### 2020 Update | ##### 2020 Update |