File size: 1,600 Bytes
68db0a3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
{
  "materials": [
    "rotated_stripes",
    "double_inclusion",
    "smooth_checker"
  ],
  "grids": [
    64,
    72,
    80,
    88,
    96,
    104,
    112,
    120,
    128
  ],
  "methods": [
    "full_stormer_verlet",
    "craig_bampton"
  ],
  "explicit_cfl_values": [
    0.25,
    0.125
  ],
  "modes_per_patch": 128,
  "reference_grid": 512,
  "alternative_reference_grid": 384,
  "time_end": 8.0,
  "samples": 641,
  "receiver_tolerances": [
    0.05,
    0.1
  ],
  "threads": 2,
  "cpu_affinity": [
    0,
    1
  ],
  "repetitions": 3,
  "candidate_order": "Material, increasing grid, Verlet CFL 0.25, Verlet CFL 0.125, Craig-Bampton m128; three consecutive independent constructions of each candidate.",
  "dimension": "Full pressure n^2; Craig-Bampton r=n+256. The time recurrence stores two full pressure arrays.",
  "norm": "Equal-weight Frobenius norm over all 641 output times including t=0, four receivers and three sources.",
  "cost": "Include assembly, independent basis construction, projection and eigensolve, propagation, receiver output; exclude reference evaluation, diagnostics and file I/O. Report the median of the three complete stage sums.",
  "purpose": "Grid-density diagnostic for fixed Craig-Bampton m128 and two fixed Stormer-Verlet CFL values. All 81 settings are remeasured in one batch. The original sparse-grid experiment is retained without modification.",
  "scope": "The envelopes compare the prescribed candidates over 5-10% receiver tolerance. They do not optimize reduced capacity, CFL, solver family, source distribution or continuum accuracy."
}