Location-dependent scaling sweep on Derecho. · Raw data (JSON)
Compset: 1850_DATM%NYF_SLND_SICE_MOM6%REGIONAL_SROF_SGLC_SWAV — machine: derecho, queue: develop — date range: 2020-01-01, 2020-01-31
Each domain is 100×100 grid points, flat 1000 m bathymetry, 10 vertical levels, run at NTASKS_OCN=20 and 40 on Derecho's develop queue (~24-day simulation). Does MOM6/CESM throughput vary by where on Earth the domain sits?
Note: DT (baroclinic timestep) is auto-selected by CrocoDash based on grid spacing — a coarser grid allows a larger CFL-stable DT, which reduces wallclock time independently of geographic location. All domains here use a matched DT=600 s so the comparison reflects location effects rather than timestep artifacts. Increasing or decreasing DT has a direct proportional effect on throughput.
| Domain | N=20 | N=40 |
|---|---|---|
| Caribbean | 76.02 | 115.30 |
| Equatorial Pacific | 71.04 | 115.01 |
| Arctic | 64.01 | 93.18 |
| Antarctic | 69.04 | 115.78 |
100x100 (0.1deg, ~11km), centered ~20N 75W, flat 1000m, nk=10
| NTASKS_OCN | Throughput (sim-years/day) | Total Run Time | OCN Cost (pe-hrs/sim-year) | Parallel efficiency (vs N=20) |
|---|---|---|---|---|
| 20 | 76.02 | 1.2 min | 40.4 | 100% |
| 40 | 115.30 | 47.2 s | 26.6 | 76% |
100x100 (0.1deg, ~11km), centered ~0N 140W, flat 1000m, nk=10
| NTASKS_OCN | Throughput (sim-years/day) | Total Run Time | OCN Cost (pe-hrs/sim-year) | Parallel efficiency (vs N=20) |
|---|---|---|---|---|
| 20 | 71.04 | 1.3 min | 43.2 | 100% |
| 40 | 115.01 | 47.3 s | 26.7 | 81% |
100x100 (10km, polar-stereo EPSG:3995), 1000x1000km cap on N pole, flat 1000m, nk=10
| NTASKS_OCN | Throughput (sim-years/day) | Total Run Time | OCN Cost (pe-hrs/sim-year) | Parallel efficiency (vs N=20) |
|---|---|---|---|---|
| 20 | 64.01 | 1.4 min | 48.0 | 100% |
| 40 | 93.18 | 58.4 s | 33.0 | 73% |
100x100 (36km, polar-stereo EPSG:3031), 3600x3600km cap on S pole (boundaries ~67-74S, widened to clear GLORYS's -80S coverage limit), flat 1000m, nk=10, DT=600s matched to other domains
| NTASKS_OCN | Throughput (sim-years/day) | Total Run Time | OCN Cost (pe-hrs/sim-year) | Parallel efficiency (vs N=20) |
|---|---|---|---|---|
| 20 | 69.04 | 1.3 min | 44.5 | 100% |
| 40 | 115.78 | 47.0 s | 26.5 | 84% |
100x100 (0.1deg), centered ~30N 75W, flat 1000m, nk=10 — queue: develop (N≤40); main (N=80)
x-axis: NTASKS_OCN
| NTASKS_OCN | Throughput (sim-years/day) | Total Run Time | OCN Cost (pe-hrs/sim-year) | Parallel efficiency (vs N=1) |
|---|---|---|---|---|
| 1 | 4.90 | 24.2 min | 627.2 | 100% |
| 2 | 13.04 | 9.1 min | 235.6 | 133% |
| 4 | 15.45 | 7.7 min | 198.8 | 79% |
| 5 | 17.91 | 6.6 min | 171.6 | 73% |
| 10 | 34.22 | 3.5 min | 89.8 | 70% |
| 20 | 69.27 | 1.7 min | 44.4 | 71% |
| 40 | 120.48 | 58.9 s | 25.5 | 61% |
| 80 | 148.46 | 47.8 s | 20.7 | 38% |