Model & assumptions
What was actually compared?
- 5,012,640 true DOFs and 9,773,946 TET4 elements.
- PETSc 3.25.1, MPICH 5.0.1, 8 ranks, AIJ and CG/GAMG.
- Both independent runs use the same input, configuration and source digests.
Historical bounded Silver evidence · two complete runs
PETSc · 5M Silver replays
Exact structured TET4 linear static workload on the frozen single host configuration. Execution source: 04b07e00ddfe0b339b5790493a607ec902b1ed80.
Scope boundaries
Where the evidence stops
- Only the exact frozen workload and environment are covered.
- No universal 5M capacity, general scaling, mixed mesh, nonlinear, GPU, restart or Gold claim.
- Source values are rounded here; exact values and provenance remain in the records.
Recorded observables
Displayed values are rounded. Exact values, threshold definitions and original execution identities remain in the source records.
| Observable | Recorded value | Frozen threshold |
|---|---|---|
| CG iterations in each run | 1,243 | Read the source contract |
| Free relative residual, run 1 | 9.8472e−11 | Read the source contract |
| Equilibrium relative error, run 1 | 1.3908e−9 | Read the source contract |
| Energy relative error, run 1 | 1.5437e−14 | Read the source contract |
| Replay verdict | PASS | Read the source contract |
Open the original records.
These links use the published source tag. The files preserve their own historical source hashes and decision status; viewing them is not a rerun.
Scope & closeout report ↗docs/verification/0_2_7/0_2_7_lu2_wp04_wp05_5m_closeout.mdSilver run 1 ↗qualification/0_2_7/wp05_runtime/wp05_5m_silver_run1.jsonSilver run 2 ↗qualification/0_2_7/wp05_runtime/wp05_5m_silver_run2.jsonReplay & configuration record ↗qualification/0_2_7/wp05_runtime/wp05_5m_silver_replay.json