Differences

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

--- namd [2020/11/20 20:30]
root
+++ namd [2022/03/14 19:10] (current)
root
@@ Line 4: / Line 4: @@
 This is for multiple-node runs with ''charmrun'' as the distributed component and ''namd2'' on each compute node. We have found most runs are faster with the ''+setcpuaffinity +isomalloc_sync'' options. charmrun ''++ppn'' should match PBS ''ppn=''.
 <code>
 module load namd/2.12  [or 2.11]
@@ Line 46: / Line 47: @@
 Intel Razor    16      2.12          1   242
 AMD Trestles   32      2.12          1   315
-Intel Pinnacle 32      2.12          1   127
+Intel G6130    32      2.12          1   127
-Intel Pinnacle 32      2.13          1   127
+Intel G6130    32      2.13          1   127
-Intel Pinnacle 32      2.15a1-AVX512 1    76
+AMD Epyc 7402  48      2.13          1    89
-Intel Pinnacle 32 V100 3.0a7-cuda    1    39
+Intel G6130    32      2.15a1-AVX512 1    76
-AMD Epyc 2     48      2.15a1-AVX2 needs recompilation
+Intel G6130    32 V100 3.0a7-cuda    1    39
+AMD Epyc 7402  48      2.15a1-AVX2 needs recompilation
 </code>

Arkansas High Performace Computing Center [hpcwiki]