OPM
Ghost entries skipped for ILU apply and SpMV operator in all levels of AMG/CPR hierarchy
Fortunately, no time stepping changes and hence failed tests. Just new warnings that need to be removed.
@blattms: I believe the last comment "benchmark please" did not have any effect - can you show me where I can check that?
turns out you can't. I asked Michael to whitelist you and @aritorto. It is "benchmark please" and should be used scarcely. We cannot really see whether it worked. Normally the benchmarking report is added to the PR after a few hours, but currently that is broken, too.
Benchmark result overview:
| Test | Configuration | Relative |
|---|---|---|
| opm-git | OPM Benchmark: drogon - Threads: 1 | 0.999 |
| opm-git | OPM Benchmark: drogon - Threads: 8 | 0.997 |
| opm-git | OPM Benchmark: punqs3 - Threads: 1 | 1.007 |
| opm-git | OPM Benchmark: punqs3 - Threads: 8 | 1.003 |
| opm-git | OPM Benchmark: smeaheia - Threads: 1 | 0.965 |
| opm-git | OPM Benchmark: smeaheia - Threads: 8 | 1 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 1 | 1.013 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 8 | 1.004 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 1 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 8 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 1 | 0.991 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 8 | 1.006 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 1 | 1.008 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 8 | 0.998 |
- Speed-up = Total time master / Total time pull request. Above 1.0 is an improvement. *
View result details @ https://www.ytelses.com/opm/?page=result&id=2370
Benchmark result overview:
| Test | Configuration | Relative |
|---|---|---|
| opm-git | OPM Benchmark: drogon - Threads: 1 | 1.005 |
| opm-git | OPM Benchmark: drogon - Threads: 8 | 0.816 |
| opm-git | OPM Benchmark: punqs3 - Threads: 1 | 1.002 |
| opm-git | OPM Benchmark: punqs3 - Threads: 8 | 0.988 |
| opm-git | OPM Benchmark: smeaheia - Threads: 1 | 0.955 |
| opm-git | OPM Benchmark: smeaheia - Threads: 8 | 0.885 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 1 | 1.008 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 8 | 0.998 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 1 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 8 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 1 | 0.99 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 8 | 0.956 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 1 | 1.001 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 8 | 0.931 |
- Speed-up = Total time master / Total time pull request. Above 1.0 is an improvement. *
View result details @ https://www.ytelses.com/opm/?page=result&id=2371
Benchmark result overview:
| Test | Configuration | Relative |
|---|---|---|
| opm-git | OPM Benchmark: drogon - Threads: 1 | 1.003 |
| opm-git | OPM Benchmark: drogon - Threads: 8 | 0.992 |
| opm-git | OPM Benchmark: punqs3 - Threads: 1 | 0.991 |
| opm-git | OPM Benchmark: punqs3 - Threads: 8 | 1.012 |
| opm-git | OPM Benchmark: smeaheia - Threads: 1 | 0.969 |
| opm-git | OPM Benchmark: smeaheia - Threads: 8 | 1.001 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 1 | 1.002 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 8 | 1.001 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 1 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 8 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 1 | 0.995 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 8 | 1.001 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 1 | 0.995 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 8 | 0.993 |
- Speed-up = Total time master / Total time pull request. Above 1.0 is an improvement. *
View result details @ https://www.ytelses.com/opm/?page=result&id=2372
Benchmark result overview:
| Test | Configuration | Relative |
|---|---|---|
| opm-git | OPM Benchmark: drogon - Threads: 1 | 0.998 |
| opm-git | OPM Benchmark: drogon - Threads: 8 | 0.992 |
| opm-git | OPM Benchmark: punqs3 - Threads: 1 | 0.989 |
| opm-git | OPM Benchmark: punqs3 - Threads: 8 | 1.009 |
| opm-git | OPM Benchmark: smeaheia - Threads: 1 | 0.969 |
| opm-git | OPM Benchmark: smeaheia - Threads: 8 | 1.001 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 1 | 1.019 |
| opm-git | OPM Benchmark: spe10_model_1 - Threads: 8 | 1.005 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 1 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_extra - Threads: 8 - FOIT (Total Oil Injection At End Of Run) | 1 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 1 | 1.006 |
| opm-git | OPM Benchmark: flow_mpi_norne - Threads: 8 | 1.005 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 1 | 0.999 |
| opm-git | OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 8 | 0.993 |
- Speed-up = Total time master / Total time pull request. Above 1.0 is an improvement. *
View result details @ https://www.ytelses.com/opm/?page=result&id=2373
Looks like no change in the benchmarks, which is as expected, since they all run using the default linear solver (i.e. ILU0 preconditioner, no CPR/AMG) I believe.
Looks like no change in the benchmarks, which is as expected, since they all run using the default linear solver (i.e. ILU0 preconditioner, no CPR/AMG) I believe.
Yes that makes sense. Curretly, I'm still waiting for Andreas to get back to me with the measurements he has done previously.
Looks like no change in the benchmarks, which is as expected, since they all run using the default linear solver (i.e. ILU0 preconditioner, no CPR/AMG) I believe.
Yes that makes sense. Curretly, I'm still waiting for Andreas to get back to me with the measurements he has done previously.
New results using the current master (N = number of processes, simulation time in seconds for normal and skipping ghost) below, for 32 and 64 processes the improvement is still around 5% and for 128 processes, the improvement is still around 10%!
| N | normal | skip ghost |
|---|---|---|
| 32 | 1043.34 | 1006.81 |
| 64 | 820.60 | 773.49 |
| 128 | 589.27 | 541.67 |
| 256 | still running | still running |
We would like to merge this. This might change parallel cpr a bit. I am not sure whether this will interfere with #5157 . Comments @steink , @GitPaean ?
I think it's preferable to merge anything in the pipeline before https://github.com/OPM/opm-simulators/pull/5157. I'm also hoping to get in a few improvements prior to https://github.com/OPM/opm-simulators/pull/5157
