Author: bugman Date: Fri May 30 11:39:57 2014 New Revision: 23610 URL: http://svn.gna.org/viewcvs/relax?rev=23610&view=rev Log: Merged revisions 23578-23579,23590-23600,23603-23609 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/trunk ................ r23578 | bugman | 2014-05-28 22:39:51 +0200 (Wed, 28 May 2014) | 5 lines Created a file selection wildcard for use in the GUI for selecting peak lists. This is used in the four user functions which read peak lists. ................ r23579 | bugman | 2014-05-28 22:42:43 +0200 (Wed, 28 May 2014) | 3 lines Changed all *.* GUI file selection wildcards to *. ................ r23590 | bugman | 2014-05-29 13:22:39 +0200 (Thu, 29 May 2014) | 3 lines Initialized merge tracking via "svnmerge" with revisions "1-23214" from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/disp_speed ................ r23591 | bugman | 2014-05-29 13:46:48 +0200 (Thu, 29 May 2014) | 199 lines Merged revisions 23216-23229 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/disp_speed ........ r23216 | tlinnet | 2014-05-18 23:36:14 +0200 (Sun, 18 May 2014) | 13 lines 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. ........ r23217 | tlinnet | 2014-05-18 23:36:17 +0200 (Sun, 18 May 2014) | 12 lines 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 ........ r23218 | tlinnet | 2014-05-19 00:51:15 +0200 (Mon, 19 May 2014) | 9 lines 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. ........ r23219 | tlinnet | 2014-05-19 00:51:18 +0200 (Mon, 19 May 2014) | 10 lines 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. ........ r23220 | tlinnet | 2014-05-19 00:51:20 +0200 (Mon, 19 May 2014) | 18 lines 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. ........ r23221 | tlinnet | 2014-05-19 01:19:02 +0200 (Mon, 19 May 2014) | 20 lines 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. ........ r23222 | tlinnet | 2014-05-19 03:20:43 +0200 (Mon, 19 May 2014) | 14 lines 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. ........ r23223 | tlinnet | 2014-05-19 03:20:46 +0200 (Mon, 19 May 2014) | 8 lines 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 ........ r23224 | tlinnet | 2014-05-19 03:20:47 +0200 (Mon, 19 May 2014) | 7 lines 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 ........ r23225 | tlinnet | 2014-05-19 03:20:49 +0200 (Mon, 19 May 2014) | 7 lines 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 ........ r23226 | tlinnet | 2014-05-19 03:20:51 +0200 (Mon, 19 May 2014) | 13 lines Speed-up of model LM63. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. Change in systemtest was: test_hansen_cpmg_data_auto_analysis 13.731s -> 9.971s test_hansen_cpmg_data_auto_analysis_r2eff 13.370s -> 9.510s test_hansen_cpmg_data_to_lm63 3.254s -> 2.080s ........ r23227 | tlinnet | 2014-05-19 03:20:53 +0200 (Mon, 19 May 2014) | 10 lines Speed-up of model IT99. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. Change in speed is: test_hansen_cpmg_data_auto_analysis 9.74s -> 8.330s test_hansen_cpmg_data_to_it99 4.928s -> 3.138s ........ r23228 | tlinnet | 2014-05-19 03:20:54 +0200 (Mon, 19 May 2014) | 3 lines Added missing imports of "numpy.array". task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. ........ r23229 | tlinnet | 2014-05-19 03:20:56 +0200 (Mon, 19 May 2014) | 8 lines Speed-up of model DPL94. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. Change in speed is: test_dpl94_data_to_dpl94 19.412s -> 4.427s ........ ................ r23592 | bugman | 2014-05-29 13:55:17 +0200 (Thu, 29 May 2014) | 18 lines Merged revisions 23239 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/disp_speed ........ r23239 | tlinnet | 2014-05-19 14:44:10 +0200 (Mon, 19 May 2014) | 10 lines Math-domain catching for model B14. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ ................ r23593 | bugman | 2014-05-29 13:56:16 +0200 (Thu, 29 May 2014) | 18 lines Merged revisions 23240 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/disp_speed ........ r23240 | tlinnet | 2014-05-19 14:44:12 +0200 (Mon, 19 May 2014) | 10 lines Math-domain catching for model CR72. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ ................ r23594 | bugman | 2014-05-29 13:59:18 +0200 (Thu, 29 May 2014) | 321 lines Merged revisions 23246-23250,23259,23268-23278,23295-23304 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/disp_speed ........ r23246 | tlinnet | 2014-05-20 01:47:30 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model: 'NS CPMG 2-site expanded'. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23247 | tlinnet | 2014-05-20 01:47:32 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model CR72. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. The skipping of test when num_points > 0, is a bad implementation. If such a case should show, it is best to catch the wrong input for the calculations. This is best done with a check before running the calculations. ........ r23248 | tlinnet | 2014-05-20 01:47:34 +0200 (Tue, 20 May 2014) | 11 lines Turned of processor.run_queue() when using calc(). task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. A weird behaviour was found when running the test-suite. Somehow, queued jobs from the last test suite was not flushed and done. Running the test-suite gave errors on almost all tests. Inspecting systemtest which only used calc() for the dispersion API, showed that no jobs was queued in the calculate() method, and the call to processor.run_queue(), was turned off, solving the problem. ........ r23249 | tlinnet | 2014-05-20 01:47:37 +0200 (Tue, 20 May 2014) | 4 lines An attempt to fix the pipe_control/minimisy.py minimise() call to processor.run_queue(). It was thought that I was modifying the address of array class(back_calc). But this does not seem to be the problem. ........ r23250 | tlinnet | 2014-05-20 02:02:19 +0200 (Tue, 20 May 2014) | 1 line Fix for variable num_points not set. ........ r23259 | tlinnet | 2014-05-20 14:55:11 +0200 (Tue, 20 May 2014) | 34 lines Reverted r23250-23248 as the correct fix was done in fixing bug 22055. bug #22055 (https://gna.org/bugs/?22055) The processor.run_queue() does not clean up in uni_processor - chain-reaction failures in the test suite. The command used was: svn merge -r23250:r23247 . ..... r23250 | tlinnet | 2014-05-20 02:02:19 +0200 (Tue, 20 May 2014) | 1 line Fix for variable num_points not set. ------------------------------------------------------------------------ r23249 | tlinnet | 2014-05-20 01:47:37 +0200 (Tue, 20 May 2014) | 4 lines An attempt to fix the pipe_control/minimisy.py minimise() call to processor.run_queue(). It was thought that I was modifying the address of array class(back_calc). But this does not seem to be the problem. ------------------------------------------------------------------------ r23248 | tlinnet | 2014-05-20 01:47:34 +0200 (Tue, 20 May 2014) | 11 lines Turned of processor.run_queue() when using calc(). task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. A weird behaviour was found when running the test-suite. Somehow, queued jobs from the last test suite was not flushed and done. Running the test-suite gave errors on almost all tests. Inspecting systemtest which only used calc() for the dispersion API, showed that no jobs was queued in the calculate() method, and the call to processor.run_queue(), was turned off, solving the problem. ..... ........ r23268 | tlinnet | 2014-05-20 22:29:35 +0200 (Tue, 20 May 2014) | 8 lines Fix for commit 23246. http://svn.gna.org/viewcvs/relax?view=revision&revision=23246 The testing taking a value to a power was performed wrong. This made systemtest Relax_disp.test_hansen_cpmg_data_auto_analysis_numeric not working. ........ r23269 | tlinnet | 2014-05-20 22:29:38 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model TSMFK01. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23270 | tlinnet | 2014-05-20 22:29:40 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model TP02. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23271 | tlinnet | 2014-05-20 22:29:41 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model TAP03. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23272 | tlinnet | 2014-05-20 22:29:43 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model DPL94. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23273 | tlinnet | 2014-05-20 22:29:45 +0200 (Tue, 20 May 2014) | 3 lines Math-domain catching for model TAP03. Another check for division with 0 inserted. ........ r23274 | tlinnet | 2014-05-20 22:29:47 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model MP05. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23275 | tlinnet | 2014-05-20 22:29:49 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model IT99. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23276 | tlinnet | 2014-05-20 22:29:50 +0200 (Tue, 20 May 2014) | 7 lines Removed class object "back_calc" being updated per time point for model LM63. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23277 | tlinnet | 2014-05-20 22:29:52 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model M61. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23278 | tlinnet | 2014-05-20 22:29:54 +0200 (Tue, 20 May 2014) | 10 lines Math-domain catching for model MMQ CR72. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. These can be found via the --numpy-raise function to the systemtests. To make the code look clean, the class object "back_calc" is no longer being updated per time point, but is updated in the relax_disp target function in one go. ........ r23295 | tlinnet | 2014-05-21 13:02:55 +0200 (Wed, 21 May 2014) | 7 lines Align math-domain catching for model CR72 with trunk implementation. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. The catching of errors have to be more careful. ........ r23296 | tlinnet | 2014-05-21 13:02:56 +0200 (Wed, 21 May 2014) | 7 lines Align math-domain catching for model DPL94 with trunk implementation. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. The catching of errors have to be more careful. ........ r23297 | tlinnet | 2014-05-21 13:02:58 +0200 (Wed, 21 May 2014) | 7 lines Align math-domain catching for model IT99 with trunk implementation. task #7793: (https://gna.org/task/?7793) Speed-up of dispersion models. This is to implement catching of math domain errors, before they occur. [... 1418 lines stripped ...]