Main Purpose:
To test the workflow and gather runtime data for estimation
Deliverable:
Validation run completed without issues ( Passing to Robin for correctness test?)
one REL on each Fault in CS-v19p5 has partially failed ( See issue #1)
runtime&CH comparison (mean value within certain mag)
(Note: data of Mag>7 is assuming the runtime would not change after we fixed the bug)
| Mag | Runtime(original) | CoreHours(original) | Runtime(CUDA) |
| >5 | 0.001 | 0.75 | 0.18 |
| >6 | 0.15 | 11 | 0.75 |
| >7 | 0.91 | 218 | 2.48 |
Issues:
- Any faults with mag > 7 will be unreadable and fail the test ( using non-uniform-grid 400 )
1a. likely caused by size and how the code tries to allocate.
(endianess calculation failed due to overflow in python, maybe the same case in the binary?)
File size are incorrect as well, likely due to the same reason.