Ticket #6417 (closed: worksforme)
Unscripted testing Model/curve fitting
Reported by: | Nick Draper | Owned by: | Raquel Alvarez Banos |
---|---|---|---|
Priority: | blocker | Milestone: | Release 3.4 |
Component: | Mantid | Keywords: | |
Cc: | Blocked By: | ||
Blocking: | Tester: | Raquel Alvarez Banos |
Description (last modified by Federico M Pouzols) (diff)
See the items at the bottom of this page http://www.mantidproject.org/Unscripted_Manual_Testing
Record in the comments what environment was used
Change History
comment:1 Changed 8 years ago by Nick Draper
- Status changed from new to closed
- Owner set to Anyone
- Resolution set to fixed
comment:2 Changed 8 years ago by Nick Draper
- Status changed from closed to verify
- Resolution fixed deleted
comment:3 Changed 8 years ago by Nick Draper
- Status changed from verify to verifying
- Tester set to Nick Draper
comment:6 Changed 6 years ago by Nick Draper
- Milestone changed from Release 2.4 to Release 3.3
- Resolution set to fixed
- Tester Nick Draper deleted
Reasurrected for R3.3 testing
comment:7 Changed 6 years ago by Nick Draper
- Status changed from closed to verify
- Resolution fixed deleted
comment:8 Changed 6 years ago by Federico M Pouzols
- Status changed from verify to verifying
- Description modified (diff)
- Tester set to Federico M Pouzols
comment:9 Changed 6 years ago by Federico M Pouzols
- Status changed from verifying to closed
Test environment: Windows 7 64bit
Windows installer 3.2.20141205.1835
No important crashes or blocker issues found.
Fitting seems to work well on single-layer plots. But, as identified in #7903, it doesn't support plots with multiple layers.
I tried several function types and they all behaved well. Sequential fitting also seems to be working well. The following minimizers behave robustly even under rather pathological conditions and non-smart parameterization: Levenberg-Marquard, Simplex, Conjugate gradients, Damping, BFGS. FABADA seems to be a bit more weak, and it will often produce errors like "Convergence NOT reached after 499 iterations. Try to set better initial values for parameters:..."
comment:14 Changed 5 years ago by Raquel Alvarez Banos
- Status changed from assigned to inprogress
- Owner changed from Anyone to Raquel Alvarez Banos
comment:15 Changed 5 years ago by Raquel Alvarez Banos
- Status changed from inprogress to verify
- Resolution set to fixed
Tested on Windows 7 64 bits, installer mantid-3.3.20150428.729
Found the following issues:
- Setting the Minimizer to Conjugate gradient (Fletcher-Reevers imp.) causes an unexpected exception. The same happens with Conjugate gradient (Polak-Ribiere imp.). Ticket #11643 has been created.
- SequentialFit dialog has an option, "Plot against log", that does not seem to be working. The output graph is empty although the table with fitting results is correctly populated. Ticket #11644 was created.
comment:16 Changed 5 years ago by Raquel Alvarez Banos
I have also noticed this issue:
- When selecting a polynomial fitting function with n=-1 (or any other negative value) an error message pops up indicating that n cannot be negative, but Mantid crashes afterwards (new ticket #11652).
comment:17 Changed 5 years ago by Nick Draper
- Status changed from verify to closed
- Resolution changed from fixed to worksforme
comment:18 Changed 5 years ago by Stuart Campbell
This ticket has been transferred to github issue 7263
Opened straight to testing with a fake owner, so anyone can test.