[aspect-devel] Aspect Timer

[Eric Heien]

I just ran a test of Aspect on Lonestar using 48 processors with the box.prm sample (using more mesh refinement) and I got the following timings:

+---------------------------------------------+------------+------------+
| Total wallclock time elapsed since start    |       429s |            |
|                                             |            |            |
| Section                         | no. calls |  wall time | % of total |
+---------------------------------+-----------+------------+------------+
| Assemble Stokes system          |       131 |      1.25s |      0.29% |
| Assemble temperature system     |       131 |      3.06s |      0.71% |
| Build Stokes preconditioner     |        30 |      1.71s |       0.4% |
| Build temperature preconditioner|       131 |     0.181s |     0.042% |
| Solve Stokes system             |       131 |       4.6s |       1.1% |
| Solve temperature system        |       131 |      1.68s |      0.39% |
| Postprocessing                  |       129 |      17.3s |         4% |
| Refine mesh structure, part 1   |        27 |     0.913s |      0.21% |
| Refine mesh structure, part 2   |        27 |      0.16s |     0.037% |
| Setup dof systems               |        28 |      1.22s |      0.28% |
+---------------------------------+-----------+------------+------------+

In case the formatting is messed up, basically it seems the total wall time (429s) is about 13x greater than the sum of the sections.  In other words, 397 seconds of run time are unaccounted for out of 429 total seconds.  Are there any sections that could potentially account for this that are not yet timed?

I know the tracers take a long time so I turned them off but I still see this discrepancy (plus they should be accounted for under "Postprocessing").  One possible hint is that it seems the "mesh" file is being altered extensively during checkpoints - perhaps this is related?

Thanks for any ideas,

-Eric

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/aspect-devel/attachments/20120503/5f483a5c/attachment.htm