Re: [bug #20821] Optimisation failure / round error on windows 64 bitRe: [bug #20821] Optimisation failure / round error on windows 64 bitI think we should close the bug. I will stay with python32 setyp for relax Best Troels Troels Emtekær Linnet 2013/6/20 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:
Hi, Sorry for all the hassles! Normally you don't need the Microsoft compiler to use relax as the C modules are pre-compiled for 32-bit Python systems (http://www.nmr-relax.com/download.html#MS_Windows). It's only relax developers who need to use this. Sometimes users will install non-supported software (64-bit Python, non-official Python, etc.) and then they may require it. But I can only support one official Python release, currently only 2.7, for the MS Windows distribution archive. Once the relax_disp branch supports more models, then it will be merged back into the relax trunk and probably released as relax 3.1. Though first relax 3.0 will need to be released. It may take a while for relax 3.1 to come out with full support for relaxation dispersion analyses, and hence pre-compiled modules. As for the errors with 64-bit WinPython, it is probably safe to assume that they have jumped the gun and bundled 64-bit Numpy for Windows before it is ready to be shipped to users. What is strange is that on my Linux test system, I have self compiled 64-bit Python+numpy+scipy+wxPython and I don't have any precision issues. There is something fishy with this WinPython 2.7.5 version. Cheers, Edward On 20 June 2013 14:09, Troels Emtekær Linnet <tlinnet@xxxxxxxxx> wrote:Hi Edward. I have spent several hours, to try to fix this issue. If you have winpython 64 bit, you need - Visual Studio Professional 2012 to compile. This is because: target_functions\relax_fit.c make an include to the header file #include <Python.h> And if you only have "Visual Studio Express 2012 for Windows Desktop", there will be a link error. I think then Scons try to compile x86, and the header files in python will not be matched. If you have professional, Scons try to compile amd64, and will find the correct link in the header files. (Or something like this ) This setup is listed here: http://nmr-relax.kimlinnet.dk/index.php?title=Installation_windows_Python_x86-64_Visual_Studio_Express_Professional But I am happy to announce, that with python 32 I have NO errors! This setup is listed here: http://nmr-relax.kimlinnet.dk/index.php?title=Installation_windows_Python_x86-32_Visual_Studio_Express_for_Windows_Desktop Weird! Troels Emtekær Linnet 2013/6/20 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Ok, then we'll have to go with that. It's good that the icons are back :) Could I ask you to make one more small change to patch3 (https://gna.org/support/download.php?file_id=18123). Could you add a line or two to the comments with information about the WinPython version that you are using, and anything else to do with WinPython that you think might be important? It would be good to make it clear in the comments that this is not results from the official Python sources. Cheers, Edward On 20 June 2013 13:11, Troels Emtekær Linnet <tlinnet@xxxxxxxxx> wrote:The te, should have places=1, since else I get the error. Traceback (most recent call last): File "C:\WinPython27\relax_disp_64\test_suite\system_tests\model_free.py", line 1782, in test_opt_constr_cd_mt_S2_0_970_te_2048_Rex_0_149 self.value_test(spin, select=select, s2=s2, te=te, rex=rex, chi2=chi2) File "C:\WinPython27\relax_disp_64\test_suite\system_tests\model_free.py", line 3142, in value_test self.assertAlmostEqual(spin.te / 1e-12, te, places=2, msg=mesg) AssertionError: Optimisation failure. And I am enjoying the icons in the menu. :-) /Troels 2013/6/20 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Does this work with places=2 or 3 for the te value? That would be better. The precision is shocking, I have to say. I would suggest that in the comments, that something about this being the WinPython distribution would also be useful for these precision records. I would not think that it is the Microsoft C compiler that is a problem, but rather the numpy packages bundled with this Python derivative. The Microsoft compiler has never been an issue before, so I would suggest that the WinPython people have a precision bug in their code. Or maybe they have deliberately turned down their precision so that it appears as if their distribution is much faster than standard Python+numpy? Is it also WinPython that has problems with icons in the menus? Regards, Edward On 20 June 2013 09:18, Troels E. Linnet <NO-REPLY.INVALID-ADDRESS@xxxxxxx> wrote:Follow-up Comment #5, bug #20821 (project relax): Reply to bug #20821: (https://gna.org/bugs/?20821) Optimisation failure / round error on windows 64 bit I experienced more precision problems. I had a terrible need, to turn down precision, quite a lot. So much, that I think something is wrong with the MS Visual Studio Express """ Professional """ 64 bit compilation. - self.assertAlmostEqual(spin.s2, s2, msg=mesg) + self.assertAlmostEqual(spin.s2, s2, places=6, msg=mesg) - self.assertAlmostEqual(spin.te / 1e-12, te, 5, msg=mesg) + self.assertAlmostEqual(spin.te / 1e-12, te, places=1, msg=mesg) - self.assertAlmostEqual(spin.rex * (2.0 * pi * cdp.spectrometer_frq[cdp.ri_ids[0]])**2, rex * (2.0 * pi * cdp.spectrometer_frq[cdp.ri_ids[0]])**2, msg=mesg) + self.assertAlmostEqual(spin.rex * (2.0 * pi * cdp.spectrometer_frq[cdp.ri_ids[0]])**2, rex * (2.0 * pi * cdp.spectrometer_frq[cdp.ri_ids[0]])**2, places=4, msg=mesg) # Optimisation values. s2 = 0.9699999999999995 te = 2048.000000000022283 rex = 0.14900000000000566 / (2.0 * pi * cdp.spectrometer_frq[cdp.ri_ids[0]])**2 My value # s2: 0.9700002183674102 : Diff -2.18367410709e-07 # te: 2048.015293187 Diff: -0.0152931869779 # rex: 0.14899473115727899 Diff: 5.26884272667e-06 # 64-bit Windows 7 (https://gna.org/bugs/?20821). # System: Windows # Release: 7 # Version: 6.1.7601 # Architecture: 64bit WindowsPE # Machine: AMD64 # Processor: Intel64 Family 6 Model 37 Stepping 2, GenuineIntel # Python version: 2.7.5 # numpy version: 1.7.1 # Compilation: Visual Studio Express 2012, Professional, for 64 bit compilation # s2: 0.9700002183674102 # te: 2048.015293187 # rex: 0.14899473115727899 # chi2: 2.3195994119090742e-10 # iter: 116 # f_count: 411 # g_count: 411 # h_count: 0 # warning: None (file #18123) _______________________________________________________ Additional Item Attachment: File name: patch3 Size:2 KB _______________________________________________________ Reply to this item at: <http://gna.org/bugs/?20821> _______________________________________________ Message sent via/by Gna! http://gna.org/ _______________________________________________ relax (http://www.nmr-relax.com) This is the relax-devel mailing list relax-devel@xxxxxxx To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-devel