relax SVN archive

Speed-up for model M61.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

Change in speed is:
test_m61_data_to_m61
6.692s -> 3.480s

Speed-up of model MMQ CR72.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

Change in systemtest:
test_sprangers_data_to_mmq_cr72
9.892s -> 4.121s

Speed-up of model MP05.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

The change in systemtest is:

test_tp02_data_to_mp05
10.750s -> 6.644s

Huge speed-up for model TAP03.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

The change for running systemtest is:
test_tp02_data_to_tap03
13.869s -> 7.263s

This is won by not checking single values in the R1rho array for math domain
errors, but calculating all steps, and in one single round check for finite values.
If just one non-finite value is found, the whole array is returned with a large
penalty of 1e100.

This makes all calculations be the fastest numpy array way.

Speed-up of model TP02.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

The change for running systemtest is:
test_curve_type_r1rho_fixed_time
0.057s -> 0.049s

test_tp02_data_to_ns_r1rho_2site
10.539s -> 10.456s

test_tp02_data_to_tp02
8.608s -> 5.727s

This is won by not checking single values in the R1rho array for math domain
errors, but calculating all steps, and in one single round check for finite values.
If just one non-finite value is found, the whole array is returned with a large
penalty of 1e100.

This makes all calculations be the fastest numpy array way.

Huge speed-up of model B14.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

Time test for systemtests:

test_baldwin_synthetic
2.626s -> 1.990s

test_baldwin_synthetic_full
18.326s -> 13.742s

This is won by not checking single values in the R2eff array for math domain
errors, but calculating all steps, and in one single round check for finite values.
If just one non-finite value is found, the whole array is returned with a large
penalty of 1e100.

This makes all calculations be the fastest numpy array way.

Speed-up of model TSMFK01.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

This is won by not checking single values in the R2eff array for math domain
errors, but calculating all steps, and in one single round check for finite values.
If just one non-finite value is found, the whole array is returned with a large
penalty of 1e100.

This makes all calculations be the fastest numpy array way.

Critical fixes for systemtest: Relax_disp.test_hansen_cpmg_data_missing_auto_analysis.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

This was found necessary after commit 23216.

It is suspected that when relax have touched boundary values which made
math domain errors, the error cathing have created local minima or
interfered with the simplex search algorithm.

Fix for systemtest test_cpmg_synthetic_dx_map_points.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

This is after commit 23216:
The systemtest uses: GRID_INC = None, which have found to be catastrophic: https://gna.org/bugs/?22032.
Commit 23216 fixed this, and now the found minimization point is much much better.

Start:
0.8 3.92 0.39964
End:
0.76982 3.9169 0.41353

Huge speed-up for model CR72.

task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models.

Systemtest Relax_disp.test_cpmg_synthetic_cr72_full_noise_cluster
changes from 7 seconds to 4.5 seconds.

This is won by not checking single values in the R2eff array for math domain
errors, but calculating all steps, and in one single round check for finite values.
If just one non-finite value is found, the whole array is returned with a large
penalty of 1e100.

This makes all calculations be the fastest numpy array way.