--- Log opened Sun Jul 08 00:00:17 2012 | ||
wiking | oh configure script why you no work! | 00:02 |
---|---|---|
blackburn | n4nd0: want to fix some bugs? | 00:03 |
blackburn | :D | 00:03 |
n4nd0 | blackburn: I am bit busy with some other bugs in the HM-SVM | 00:04 |
n4nd0 | but yeah of course | 00:04 |
blackburn | ah ok | 00:04 |
n4nd0 | I can lend a hand there if so | 00:04 |
blackburn | no rather do HM-SVM :) | 00:04 |
n4nd0 | where are they? | 00:04 |
blackburn | n4nd0: mainly in serialization and parameter trees | 00:06 |
blackburn | good news for heiko | 00:06 |
blackburn | :D | 00:06 |
n4nd0 | :S | 00:06 |
n4nd0 | files? where do they fire? | 00:06 |
blackburn | ah | 00:07 |
blackburn | serialization_basic_tests | 00:07 |
blackburn | ==15481== HEAP SUMMARY: | 00:07 |
blackburn | ==15481== LEAK SUMMARY: | 00:07 |
blackburn | ==15481== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 2 from 2) | 00:07 |
blackburn | base_load_file_parameters | 00:07 |
blackburn | ==15235== HEAP SUMMARY: | 00:07 |
blackburn | ==15235== LEAK SUMMARY: | 00:07 |
blackburn | ==15235== ERROR SUMMARY: 3 errors from 3 contexts (suppressed: 2 from 2) | 00:07 |
blackburn | base_load_all_file_parameters | 00:07 |
blackburn | ==15241== HEAP SUMMARY: | 00:07 |
blackburn | ==15241== LEAK SUMMARY: | 00:07 |
blackburn | ==15241== ERROR SUMMARY: 3 errors from 3 contexts (suppressed: 2 from 2) | 00:08 |
blackburn | base_map_parameters | 00:08 |
blackburn | ==15247== HEAP SUMMARY: | 00:08 |
blackburn | ==15247== LEAK SUMMARY: | 00:08 |
blackburn | ==15247== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 2 from 2) | 00:08 |
blackburn | bahh | 00:08 |
n4nd0 | ok | 00:11 |
blackburn | n4nd0: nevermind let heiko fix that :) | 00:12 |
n4nd0 | I will not start with them right away because I am with this HM-SVM test | 00:12 |
blackburn | no problem | 00:12 |
n4nd0 | blackburn: I will tell you again once I am done and if they are not fixed I will dive in ;) | 00:12 |
wiking | anyone seen something like this before | 00:16 |
wiking | features/SparseFeatures.cpp: In member function 'void shogun::CSparseFeatures<ST>::free_sparse_feature_matrix()': | 00:16 |
wiking | features/SparseFeatures.cpp:95: error: expected class-name before '(' token | 00:16 |
blackburn | wiking: wow | 00:17 |
n4nd0 | wiking: I don't recognize it | 00:18 |
wiking | although i'm a bit puzzled by that line | 00:18 |
wiking | SGSparseVector<ST>* sparse_feature_matrix; | 00:18 |
blackburn | wiking: try to add include of SGSparseVector | 00:18 |
wiking | aah ok | 00:18 |
wiking | (&sparse_feature_matrix[i])->~SGSparseVector(); | 00:18 |
blackburn | just some guess | 00:18 |
blackburn | don't ask :D | 00:19 |
blackburn | well | 00:19 |
wiking | why do we need the address? | 00:19 |
blackburn | it is caused | 00:19 |
wiking | and not just sparse_feature_matrix[].~ ... | 00:19 |
blackburn | by the allocate thing | 00:19 |
blackburn | hmm | 00:19 |
blackburn | yes you are right it is the same | 00:19 |
blackburn | I thought you are asking why the heck we are calling dtor manually | 00:20 |
wiking | mmm | 00:20 |
wiking | but it doesn't help :)) | 00:20 |
blackburn | wiking: include? | 00:20 |
blackburn | interesting | 00:21 |
wiking | mmm i'll try now | 00:21 |
blackburn | wiking: what doesn't help? | 00:22 |
wiking | nop | 00:22 |
blackburn | prediction error on training data (3 classes): 0.308333 random guess error would be: 0.666667 | 00:23 |
blackburn | LOL | 00:23 |
wiking | ((SGSparseVector<ST>*)&sparse_feature_matrix[i])->~SGSparseVector(); | 00:24 |
wiking | doesn't help either | 00:24 |
wiking | hahah this is weird shit | 00:26 |
blackburn | yeah kind of | 00:26 |
wiking | mmm let's see | 00:27 |
wiking | ok | 00:27 |
wiking | it's the gcc | 00:27 |
blackburn | wiking: what is -v? | 00:27 |
wiking | c++ (GCC) 4.2.1 20070831 patched [FreeBSD] | 00:28 |
wiking | so this doesn't like it | 00:28 |
blackburn | woooow | 00:28 |
blackburn | :D | 00:28 |
wiking | c++46 (FreeBSD Ports Collection) 4.6.4 20120608 (prerelease) | 00:28 |
wiking | takes it | 00:28 |
blackburn | newer than mine | 00:28 |
wiking | what's your numba? | 00:28 |
blackburn | gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) | 00:29 |
wiking | anybody has 4.2.x compiler around? | 00:29 |
wiking | on a different os? | 00:29 |
blackburn | IIRC 4.2 is really some crappy v | 00:30 |
blackburn | or <4.4 | 00:30 |
wiking | yeah but still | 00:30 |
wiking | we need to check then in configure script ;P | 00:30 |
wiking | ok | 00:35 |
wiking | same with 4.2 c++ on osx | 00:35 |
blackburn | wiking: does it work with 4.6 on your machine? | 00:36 |
wiking | blackburn: on fbsd it does with 4.6 | 00:37 |
wiking | on osx i have clang 3.1 by default | 00:37 |
wiking | and that does as well | 00:37 |
wiking | only that 4.2.1 does not like it neither on osx nor on freebsd | 00:38 |
wiking | hahaha | 00:40 |
wiking | i think it's a gcc bug | 00:40 |
blackburn | yeah seems to be | 00:41 |
wiking | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23287 | 00:41 |
wiking | here's the bug for it | 00:41 |
wiking | :) | 00:41 |
blackburn | argh | 00:41 |
wiking | Fixed for 4.4. | 00:42 |
wiking | fixed for 4.3 as well, closing. | 00:42 |
wiking | ok | 00:42 |
wiking | so this is crashing in 4.1.x and 4.2.x | 00:42 |
wiking | only works from 4.3.x | 00:42 |
wiking | so i guess it's safe to add a gcc 4.3.x dependency in the configure script | 00:42 |
blackburn | wiking: yeah | 00:45 |
CIA-18 | shogun: Sergey Lisitsyn master * r77ab132 / (3 files in 3 dirs): Fixed MKL libshogun example - http://git.io/bK_ZRA | 00:45 |
blackburn | wiking: do we check version? | 00:46 |
wiking | yep | 00:46 |
blackburn | wiking: where? | 00:47 |
wiking | case $cxx_version in | 00:47 |
wiking | 2.96*) | 00:47 |
wiking | cxx_version="$cxx_version, bad" | 00:47 |
wiking | cxx_found=no | 00:47 |
wiking | apparently 2.96* is bad :) | 00:47 |
shogun-buildbot | build #83 of deb1 - libshogun started, including [77ab1329c966b952c91e09c0c1bf7a48f29fa967] | 00:47 |
wiking | everything else is good | 00:47 |
wiking | :>>> | 00:47 |
shogun-buildbot | build #83 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/83 | 00:48 |
shogun-buildbot | build #89 of deb2 - static_interfaces started, including [77ab1329c966b952c91e09c0c1bf7a48f29fa967] | 00:48 |
shogun-buildbot | build #91 of deb3 - modular_interfaces started, including [77ab1329c966b952c91e09c0c1bf7a48f29fa967] | 00:48 |
shogun-buildbot | build #11 of cyg1 - libshogun started, including [77ab1329c966b952c91e09c0c1bf7a48f29fa967] | 00:48 |
blackburn | we seriously need to clean up configure | 00:49 |
shogun-buildbot | build #89 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/89 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 00:49 |
n4nd0 | good night men of the night's watch, it's time for me to retire :) | 00:53 |
wiking | gnite | 00:53 |
blackburn | nite | 00:53 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 00:54 | |
CIA-18 | shogun: Sergey Lisitsyn master * rb32d3e0 / (2 files in 2 dirs): Fixed kernel and bmrm warnings - http://git.io/3UImvg | 00:55 |
shogun-buildbot | build #84 of deb1 - libshogun started, including [b32d3e011bc5892f929166beb450aae09366f1a1] | 00:57 |
shogun-buildbot | build #84 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/84 | 00:57 |
shogun-buildbot | build #92 of deb3 - modular_interfaces started, including [b32d3e011bc5892f929166beb450aae09366f1a1] | 00:57 |
wiking | ok | 01:00 |
wiking | i just need to figure it out now with the bash case | 01:00 |
CIA-18 | shogun: Sergey Lisitsyn master * rbfa3239 / src/shogun/classifier/svm/OnlineLibLinear.h : Made OnlineLibLinear fields private - http://git.io/306ylQ | 01:00 |
CIA-18 | shogun: Sergey Lisitsyn master * rbe76f99 / (6 files in 3 dirs): Revert "Added PegasosSVM" - http://git.io/QfqmRQ | 01:00 |
shogun-buildbot | build #11 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/11 | 01:00 |
shogun-buildbot | build #12 of cyg1 - libshogun started, including [b32d3e011bc5892f929166beb450aae09366f1a1] | 01:00 |
wiking | and then actually we have shogun compiling on fbsd | 01:01 |
* wiking wonders why is it so fucking slow... | 01:04 | |
wiking | another magic: error: '::getline' has not been declared | 01:07 |
blackburn | :D | 01:08 |
shogun-buildbot | build #91 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/91 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:08 |
shogun-buildbot | build #90 of deb2 - static_interfaces started, including [b32d3e011bc5892f929166beb450aae09366f1a1] | 01:08 |
blackburn | where? | 01:08 |
wiking | io/AsciiFile.cpp | 01:08 |
blackburn | nice | 01:08 |
blackburn | wiking: first googling shows that it probably broken in freebsd's gcc :D | 01:09 |
wiking | no fucking way | 01:09 |
wiking | it does have getline | 01:10 |
wiking | but not ::getline | 01:10 |
wiking | aah ok | 01:11 |
wiking | i see from the man | 01:11 |
wiking | ;) | 01:11 |
blackburn | wiking: what do you see? | 01:11 |
wiking | defining the macro _WITH_GETLINE, prior to the inclusion of <stdio.h>. | 01:11 |
blackburn | wiking: I can do this and push, or will you do? | 01:12 |
shogun-buildbot | build #92 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/92 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:12 |
shogun-buildbot | build #85 of deb1 - libshogun started, including [bfa3239987a7831f96945aa69c158adbfff65609, be76f9993dfb8ae69582a02bfb3891ec917ad61f] | 01:12 |
shogun-buildbot | build #12 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/12 | 01:12 |
wiking | blackburn: didn't help yet | 01:12 |
blackburn | hmm | 01:12 |
wiking | blackburn: i'll push the PR when i have all the fixes for fbsd | 01:12 |
blackburn | thanks | 01:12 |
shogun-buildbot | build #85 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/85 | 01:13 |
shogun-buildbot | build #13 of cyg1 - libshogun started, including [bfa3239987a7831f96945aa69c158adbfff65609] | 01:13 |
shogun-buildbot | build #93 of deb3 - modular_interfaces started, including [bfa3239987a7831f96945aa69c158adbfff65609, be76f9993dfb8ae69582a02bfb3891ec917ad61f] | 01:13 |
shogun-buildbot | build #90 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/90 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:13 |
shogun-buildbot | build #91 of deb2 - static_interfaces started, including [bfa3239987a7831f96945aa69c158adbfff65609, be76f9993dfb8ae69582a02bfb3891ec917ad61f] | 01:13 |
wiking | i did it as they suggest but no help | 01:13 |
wiking | only if i remove the :: | 01:14 |
blackburn | wiking: ifdef it when | 01:14 |
wiking | and then it works even w/o the define | 01:14 |
wiking | yeah i guess that'll be it | 01:14 |
wiking | just lets see how it ends | 01:14 |
wiking | since it's still compiling | 01:14 |
wiking | done | 01:16 |
wiking | ok that actually wasn't that bad... | 01:16 |
wiking | only 2 file needs to be changed | 01:16 |
shogun-buildbot | build #91 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/91 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:18 |
blackburn | wiking: do you know wheter clang supports openmp? | 01:22 |
shogun-buildbot | build #13 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/13 | 01:23 |
shogun-buildbot | build #14 of cyg1 - libshogun started, including [be76f9993dfb8ae69582a02bfb3891ec917ad61f] | 01:23 |
wiking | blackburn: yeps | 01:25 |
wiking | it does | 01:25 |
wiking | although it is from 3.2 i think | 01:26 |
wiking | :) | 01:26 |
wiking | https://github.com/motonacciu/clomp | 01:26 |
blackburn | argh | 01:27 |
blackburn | so we shall wait | 01:27 |
wiking | http://polly.llvm.org/ | 01:27 |
wiking | :) | 01:27 |
wiking | http://polly.llvm.org/performance.html | 01:30 |
wiking | looks good | 01:30 |
wiking | :) | 01:30 |
shogun-buildbot | build #93 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/93 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 01:32 |
shogun-buildbot | build #14 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/14 | 01:36 |
CIA-18 | shogun: Sergey Lisitsyn master * rf3871d0 / (2 files in 2 dirs): Made parameters work for SLEP FBLR - http://git.io/tyAciQ | 02:12 |
shogun-buildbot | build #86 of deb1 - libshogun started, including [f3871d06c0c06f3bfe5c9d8550110ec283038014] | 02:17 |
shogun-buildbot | build #86 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/86 | 02:21 |
shogun-buildbot | build #94 of deb3 - modular_interfaces started, including [f3871d06c0c06f3bfe5c9d8550110ec283038014] | 02:21 |
shogun-buildbot | build #92 of deb2 - static_interfaces started, including [f3871d06c0c06f3bfe5c9d8550110ec283038014] | 02:21 |
shogun-buildbot | build #15 of cyg1 - libshogun started, including [f3871d06c0c06f3bfe5c9d8550110ec283038014] | 02:21 |
shogun-buildbot | build #92 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/92 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:27 |
shogun-buildbot | build #15 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/15 | 02:33 |
shogun-buildbot | build #94 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/94 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 02:36 |
CIA-18 | shogun: Sergey Lisitsyn master * ra838aaa / (2 files): Made print result virtual and available in EvaluationResult - http://git.io/FXjBtg | 02:51 |
shogun-buildbot | build #87 of deb1 - libshogun started, including [a838aaa11a3cc989e453eb17aa57d7ce8bdd0da1] | 02:52 |
-!- blackburn [~blackburn@81.28.187.0] has left #shogun [] | 02:54 | |
shogun-buildbot | build #87 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/87 | 02:56 |
shogun-buildbot | build #93 of deb2 - static_interfaces started, including [a838aaa11a3cc989e453eb17aa57d7ce8bdd0da1] | 02:56 |
shogun-buildbot | build #95 of deb3 - modular_interfaces started, including [a838aaa11a3cc989e453eb17aa57d7ce8bdd0da1] | 02:56 |
shogun-buildbot | build #16 of cyg1 - libshogun started, including [a838aaa11a3cc989e453eb17aa57d7ce8bdd0da1] | 02:56 |
shogun-buildbot | build #93 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/93 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 03:02 |
shogun-buildbot | build #11 of nightly_all started, including [] | 03:02 |
shogun-buildbot | build #16 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/16 | 03:08 |
@sonney2k | rieck, hey konrad can we have betty back? | 03:08 |
shogun-buildbot | build #95 of deb3 - modular_interfaces is complete: Failure [failed compile csharp_modular] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/95 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 03:15 |
shogun-buildbot | build #16 of nightly_default started, including [] | 03:15 |
shogun-buildbot | build #11 of nightly_all is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/11 | 03:19 |
shogun-buildbot | build #14 of nightly_none started, including [] | 03:19 |
shogun-buildbot | build #14 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/14 | 03:27 |
CIA-18 | shogun: Soeren Sonnenburg master * r0746d4a / src/interfaces/modular/modshogun_ignores.i : ignore constructor with two SGVector arguments in csharp interface - http://git.io/d7gqgw | 03:36 |
shogun-buildbot | build #88 of deb1 - libshogun started, including [0746d4a052bfb55fd339e37d20bbb7df628b2c49] | 03:37 |
shogun-buildbot | build #16 of nightly_default is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/16 | 03:37 |
shogun-buildbot | build #88 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/88 | 03:42 |
shogun-buildbot | build #94 of deb2 - static_interfaces started, including [0746d4a052bfb55fd339e37d20bbb7df628b2c49] | 03:42 |
shogun-buildbot | build #96 of deb3 - modular_interfaces started, including [0746d4a052bfb55fd339e37d20bbb7df628b2c49] | 03:42 |
shogun-buildbot | build #17 of cyg1 - libshogun started, including [0746d4a052bfb55fd339e37d20bbb7df628b2c49] | 03:42 |
shogun-buildbot | build #94 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/94 blamelist: Soeren Sonnenburg <sonne@debian.org> | 03:47 |
shogun-buildbot | build #17 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/17 | 03:53 |
shogun-buildbot | build #96 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/96 blamelist: Soeren Sonnenburg <sonne@debian.org> | 04:08 |
-!- pluskid [~pluskid@111.120.50.40] has joined #shogun | 07:50 | |
-!- pluskid [~pluskid@111.120.50.40] has quit [Quit: Leaving] | 09:28 | |
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun | 11:44 | |
-!- blackburn [~blackburn@81.28.187.0] has quit [Quit: Leaving.] | 12:22 | |
-!- blackburn1 [~blackburn@81.28.187.0] has joined #shogun | 12:27 | |
blackburn1 | sonney2k: I solved AC issue for a while with hardresetting notebook :D | 12:28 |
blackburn1 | linux way | 12:28 |
-!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has joined #shogun | 12:40 | |
blackburn1 | heiko: I have cool news for ya | 12:45 |
blackburn1 | almost all serialization and modelselection examples have memory leaks :) | 12:48 |
heiko | blackburn1, yeeeeeah :) | 12:52 |
heiko | the did not have before | 12:52 |
heiko | argh! :) | 12:52 |
heiko | I will have a look later today, have to leave now | 12:52 |
-!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has quit [Quit: Leaving.] | 12:53 | |
-!- blackburn1 [~blackburn@81.28.187.0] has quit [Quit: Leaving.] | 12:58 | |
wiking | sonney2k: here? | 13:01 |
@sonney2k | wiking, ? | 14:22 |
@sonney2k | ...I will certainly be areound in the evening | 14:23 |
wiking | ok | 14:36 |
wiking | i'll just commit stuff and you'll comment them there ;) | 14:36 |
wiking | __MACH__ | 14:42 |
wiking | what's this macro? | 14:42 |
wiking | lol i just love http://defunkt.io/hub/ | 15:04 |
wiking | i've just sent my PR for freebsd fixes via command line ;) | 15:04 |
wiking | sonney2k: once the freebsd fixes are merge i'll need the master name and passwd for the freebsd buildslave | 16:44 |
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun | 18:12 | |
-!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has joined #shogun | 18:29 | |
@sonney2k | blackburn, google-chrome works again :-) The problem only happens when I use the kernel 3.5-rcX but not under 3.4.X ! | 18:38 |
@sonney2k | kind of wtf! | 18:38 |
blackburn | hah | 18:38 |
blackburn | kind of regression I'd say | 18:38 |
@sonney2k | I cannot tell how happy I am | 18:38 |
@sonney2k | yeah I reported that throuh all possible channels... lets see if they will laugh at me :D | 18:38 |
blackburn | sonney2k: my AC thing is solved too (for now) | 18:39 |
@sonney2k | one happy family | 18:45 |
@sonney2k | now only heiko needs to fix the build and rieck setup betty :D | 18:45 |
heiko | sonney2k, just started looking at it :) | 18:46 |
blackburn | sonney2k: do you have any overheat problems with your MBP? | 18:46 |
heiko | sonney2k, blackburn, the x-val example did work, what was recently changed? | 18:47 |
@sonney2k | blackburn, well 60C is the CPU | 18:47 |
blackburn | heiko: puffin444's stuff I guess | 18:47 |
blackburn | sonney2k: idle? | 18:48 |
@sonney2k | yeah | 18:48 |
heiko | mmh | 18:48 |
heiko | k | 18:48 |
@sonney2k | wiking, about the freebsd patch | 18:48 |
blackburn | sonney2k: I just wanted to say I managed to slow down internal gpu and it is something like -2C to before | 18:48 |
@sonney2k | wiking, are you sure that HAVE_ATLAS='#define HAVE_ATLAS 1' is correct? | 18:48 |
@sonney2k | wiking, sounded like it should be just HAVE_LAPACK='#define HAVE_LAPACK 1' | 18:49 |
@sonney2k | heiko, well the buildbot has it all - basically only puffins stuff touches the x-val stuff | 18:49 |
heiko | yeah ok, I will have a look now | 18:50 |
@sonney2k | wiking, and I don't understand the getline change | 18:50 |
@sonney2k | it must be ::getline I am afraid | 18:50 |
blackburn | sonney2k: freebsd has getline issues | 18:50 |
blackburn | :D | 18:50 |
@sonney2k | otherwise it is an endless recursion | 18:50 |
wiking | sonney2k: i was compiling with ATLAS | 18:50 |
wiking | sonney2k: so it should have HAVE_ATLAS='#define HAVE_ATLAS 1' | 18:51 |
@sonney2k | ok | 18:51 |
wiking | and as well as lapack | 18:51 |
blackburn | heiko: ah btw I made print_result of CVResult virtual | 18:51 |
blackburn | have you seen? | 18:51 |
wiking | sonney2k: for the ::getline it just says no such function... only getline... | 18:51 |
heiko | not yet, but makes sense right :) | 18:51 |
wiking | sonney2k: i couldn't fix it other way :( | 18:51 |
@sonney2k | wiking, but then it will take the one from CAsciiFile | 18:51 |
blackburn | heiko: I had troubles with printing out result | 18:51 |
blackburn | heiko: ah btw two questions | 18:52 |
heiko | k | 18:52 |
blackburn | what is ratio to divide train/test set in StratifiedCrossValidation? | 18:52 |
blackburn | how much vectors will it take as train vectors? | 18:52 |
@sonney2k | wiking, http://www.unix.com/man-page/freebsd/3/getline/ | 18:52 |
@sonney2k | seems like it works slightly differently | 18:52 |
heiko | what do you mean? | 18:53 |
heiko | mmh well you specify a number of folds | 18:53 |
wiking | sonney2k: it's the same function arg as far as i can see | 18:53 |
blackburn | heiko: right | 18:53 |
heiko | then nr_examples / folds is the size of test data | 18:53 |
blackburn | ahh | 18:53 |
blackburn | I didn't know that | 18:53 |
heiko | and one of the training sets might be larger smaller by one | 18:53 |
blackburn | heiko: is it something common practice? | 18:53 |
blackburn | I am pretty lame with xval to be honest | 18:53 |
heiko | maybe it should be specified in the doc | 18:54 |
heiko | blackburn, yes thats practice | 18:54 |
blackburn | I see | 18:54 |
heiko | divide into equal sized sets | 18:54 |
wiking | sonney2k: the only thing is this stupid hack about #define _WITH_GETLINE | 18:54 |
heiko | and then use all but one for training | 18:54 |
heiko | for all combinations | 18:54 |
wiking | sonney2k: but yeah even using that doesn't help with ::getline | 18:54 |
@sonney2k | wiking, yeah but then ::getline() should work | 18:54 |
blackburn | heiko: I thought of 50/50 or so | 18:54 |
blackburn | :D | 18:54 |
heiko | in stratified also, the class ratios are tried to kept equal | 18:54 |
wiking | sonney2k: but it doesnt :) | 18:54 |
heiko | blackburn, thats 2 folds :) | 18:54 |
blackburn | I see | 18:54 |
@sonney2k | wiking, what exactly does it say? | 18:55 |
wiking | sonney2k: lemme reproduce it | 18:55 |
blackburn | heiko: if I have one parameter only | 18:55 |
@sonney2k | wiking, with the rest I am fine | 18:55 |
blackburn | should I put it into root or to child of root? | 18:55 |
heiko | blackburn, the extreme is LOO, that is as many folds as training examples | 18:55 |
wiking | sonney2k: have u seen the gcc bug | 18:55 |
wiking | sonney2k: that's pretty funny ;) | 18:55 |
heiko | to child of root | 18:56 |
wiking | sonney2k: io/AsciiFile.cpp:1089:9: error: '::getline' has not been declared | 18:56 |
CIA-18 | shogun: Viktor Gal master * re2ce7e7 / (7 files in 4 dirs): Fix indentation to comply with shogun standard - http://git.io/xMgbeA | 18:56 |
CIA-18 | shogun: Soeren Sonnenburg master * ree792cd / (7 files in 4 dirs): Merge pull request #630 from vigsterkr/latent - http://git.io/JT_ALA | 18:56 |
shogun-buildbot | build #89 of deb1 - libshogun started, including [e2ce7e783ee19cecfaabfc1e0262b2ad1fae9772] | 18:57 |
shogun-buildbot | build #90 of deb1 - libshogun started, including [ee792cd1b55347ae3a36c1d5d8a6cf6c3f337596] | 18:57 |
blackburn | heiko: I actually have something strange with xval | 18:57 |
heiko | blackburn, what is it? | 18:57 |
blackburn | heiko: it chooses value with AUC = 0.5 :D | 18:57 |
wiking | sonney2k: and i've tried googling about this, but no help until now :( | 18:57 |
wiking | i mean about ::getline | 18:57 |
heiko | you mean model selection | 18:57 |
heiko | grid-search? | 18:58 |
blackburn | heiko: yeah sure | 18:58 |
shogun-buildbot | build #89 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/89 | 18:58 |
shogun-buildbot | build #95 of deb2 - static_interfaces started, including [e2ce7e783ee19cecfaabfc1e0262b2ad1fae9772] | 18:58 |
shogun-buildbot | build #18 of cyg1 - libshogun started, including [e2ce7e783ee19cecfaabfc1e0262b2ad1fae9772] | 18:58 |
blackburn | I do not separate these two parts so much :D | 18:58 |
heiko | it depends on which criterion you optimize, do you optimize for AUC? (Thats not always a good idea) | 18:58 |
wiking | sonney2k: another possibility is that i use that work around for freebsd as in case for __MACH__ | 18:58 |
heiko | and the direction is important, might be that you minimize or so :) | 18:59 |
@sonney2k | wiking, problem really is that this won't work with just getline() | 18:59 |
@sonney2k | just because the signatures overlap | 18:59 |
heiko | since 0.5 AUC is the worst classifier that you can construct for symmetric class ratio :) | 18:59 |
wiking | sonney2k: yeah i guess it'll be a infinite recurisve call | 18:59 |
blackburn | heiko: yeah I know | 18:59 |
blackburn | heiko: but isn't maximize default for ROC? | 18:59 |
heiko | Should be | 18:59 |
blackburn | ok I'll check | 18:59 |
heiko | but this would mean that ALL AUC values are 0.5 | 18:59 |
shogun-buildbot | build #95 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/95 blamelist: Viktor Gal <viktor.gal@maeth.com> | 19:00 |
shogun-buildbot | build #97 of deb3 - modular_interfaces started, including [e2ce7e783ee19cecfaabfc1e0262b2ad1fae9772] | 19:00 |
blackburn | heiko: yes it is easy to get with classifier I work with | 19:00 |
wiking | sonney2k: we are trying to call std::getline right? | 19:00 |
blackburn | tree lasso that is - makes w=0 at some pointt | 19:00 |
wiking | mm no... | 19:01 |
wiking | well then i really wonder | 19:01 |
heiko | blackburn, no idea of that | 19:01 |
blackburn | heiko: I will check if direction is set correct | 19:01 |
shogun-buildbot | build #90 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/90 | 19:01 |
shogun-buildbot | build #96 of deb2 - static_interfaces started, including [ee792cd1b55347ae3a36c1d5d8a6cf6c3f337596] | 19:01 |
wiking | blame meeee | 19:02 |
heiko | if not, you could add debug outputs for the results of each fold | 19:02 |
heiko | then youll see if there is something wrong | 19:02 |
blackburn | heiko: what do I see is | 19:02 |
blackburn | yes I enabled it before | 19:02 |
blackburn | AUC is better on other values | 19:02 |
heiko | you mean you optimize another criterion? | 19:03 |
heiko | and then compute AUC for result? | 19:03 |
wiking | oh ok this is not my fault ;) | 19:03 |
heiko | I remember that there were some errors in the evaluation classes some time ago, but I did not use all of them so maybe some of them are still in there | 19:03 |
heiko | blackburn, this was due to the changes in the labels | 19:03 |
heiko | remember? | 19:03 |
heiko | I never use AUC, so perhaps there is a problem or so | 19:03 |
heiko | with the transition to Binary/blabla labels | 19:04 |
blackburn | heiko: https://gist.github.com/3071792 | 19:04 |
@sonney2k | wiking, how about the following: you copy the freebsd code for getline into CAsciiFile::getline in the #ifdef __MACH__ | 19:05 |
wiking | sonney2k: or just #ifdef __MACH__ || FREEBSD | 19:05 |
wiking | ? | 19:05 |
@sonney2k | and then change it to #if defined(__MACH__) || defined(FREEBSD) | 19:05 |
heiko | blackburn that seems correct | 19:05 |
@sonney2k | wiking, the getline function is wrong in MACH It seems | 19:06 |
heiko | could you give me the data? | 19:06 |
heiko | then Ill have a look | 19:06 |
blackburn | heiko: sure | 19:06 |
@sonney2k | or maybe not | 19:06 |
heiko | blackburn, might take a while, currently fixing other bugs | 19:06 |
@sonney2k | it is actually correct :D | 19:06 |
blackburn | heiko: let me check it now then :) | 19:06 |
@sonney2k | wiking, so yes just #if defined(__MACH__) || defined(FREEBSD) | 19:06 |
blackburn | first of all I am going to check whether optimization direction is correct | 19:06 |
wiking | sonney2k: ok | 19:07 |
shogun-buildbot | build #96 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/96 blamelist: Soeren Sonnenburg <sonne@debian.org> | 19:07 |
shogun-buildbot | build #98 of deb3 - modular_interfaces started, including [ee792cd1b55347ae3a36c1d5d8a6cf6c3f337596] | 19:07 |
wiking | sonney2k: ok compiles fine... i'll push the change in a sec | 19:08 |
@sonney2k | wiking, great :D | 19:09 |
@sonney2k | heiko, please fix this http://shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/90/steps/compile/logs/warnings%20%282%29 :) | 19:09 |
shogun-buildbot | build #18 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/18 | 19:09 |
shogun-buildbot | build #19 of cyg1 - libshogun started, including [ee792cd1b55347ae3a36c1d5d8a6cf6c3f337596] | 19:09 |
heiko | sonney2k, I will, does it block something? | 19:09 |
wiking | sonney2k: ok it's there | 19:11 |
wiking | sonney2k: now let's set up the buildslave | 19:11 |
@sonney2k | heiko, nope - but easy to fix | 19:11 |
heiko | sonney2k, ok | 19:11 |
CIA-18 | shogun: Viktor Gal master * r7b3ee5a / src/configure : Fix configure script for FreeBSD - http://git.io/0s6MxA | 19:12 |
CIA-18 | shogun: Viktor Gal master * r97763bb / src/configure : gcc compiler version check - http://git.io/PG_hyw | 19:12 |
CIA-18 | shogun: Viktor Gal master * r69d4274 / src/shogun/io/AsciiFile.cpp : Fix getline function definition for FreeBSD - http://git.io/hXGOeg | 19:12 |
CIA-18 | shogun: Viktor Gal master * r643ee74 / src/shogun/io/AsciiFile.cpp : True fix for getline under FreeBSD - http://git.io/YB1YDw | 19:12 |
CIA-18 | shogun: Soeren Sonnenburg master * r28c06ce / (src/configure src/shogun/io/AsciiFile.cpp): Merge pull request #629 from vigsterkr/fbsd - http://git.io/kaC6xw | 19:12 |
wiking | letssee if it compiles alright ;) | 19:12 |
@sonney2k | wiking, hmmhh for these you should next time really use git commit --amend | 19:13 |
@sonney2k | so that it is just one commit instead of 4 | 19:13 |
@sonney2k | (and the push with --force thoug) | 19:14 |
@sonney2k | h | 19:14 |
@sonney2k | wiking, alright buildslave: | 19:14 |
@sonney2k | wiking, did you already run buildslave create-slave | 19:14 |
wiking | tried | 19:15 |
@sonney2k | wiking, send me some password you like | 19:15 |
wiking | it told me that i need to give more info | 19:15 |
wiking | basedir> <master> <name> <passwd> | 19:15 |
wiking | i guess apart from master i can give whatever i want ;) | 19:16 |
@sonney2k | lets call it shogun-freebsd as user | 19:16 |
@sonney2k | password - your choice :D | 19:17 |
blackburn | heiko: ok Z is set corrrect | 19:17 |
wiking | and what's the master? | 19:17 |
blackburn | heiko: but print_result produces something wrong I guess | 19:17 |
@sonney2k | wiking, www.shogun-toolbox.org | 19:17 |
wiking | ok | 19:18 |
@sonney2k | wiking, alright so now what do we want to build on that machine? | 19:19 |
@sonney2k | just libshogun or other stuff too? | 19:19 |
wiking | sonney2k: mmm i haven't tried other stuff | 19:19 |
@sonney2k | wiking, I would suggest to start with libshogun... | 19:19 |
heiko | blackburn, so the model selection works but the printing is wrong | 19:19 |
wiking | thus i guess i dont have the dependencies installed for everything | 19:20 |
blackburn | heiko: I guess so | 19:20 |
heiko | you could do a | 19:20 |
heiko | best_parameters.print() | 19:20 |
heiko | or print_tree or so | 19:20 |
heiko | forgot the name | 19:20 |
wiking | sonney2k: ok i've edited the files under ./info | 19:20 |
shogun-buildbot | build #19 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/19 | 19:21 |
wiking | sonney2k: so i guess it's time to start the buildbot for me? | 19:21 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 19:22 | |
wiking | sonney2k: ah yeah 2 things! 1) configure script should be run like this: ./configure --cc=gcc46 --cxx=c++46 | 19:23 |
wiking | 2) instead of make it should call gmake | 19:23 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 19:23 | |
shogun-buildbot | build #91 of deb1 - libshogun started, including [7b3ee5ae63d2dcde25cc70e4074d75ecd11b0f98, 97763bbc134688dc412b6c735fa076d85e8448b9, 69d42744dff01bf9208151d829fa7e204926227d, 643ee74dc7f43c93dc27c7be62c7349079dc17b5, 28c06ce21772e79529bf3caa39f8b484bae0dcff] | 19:23 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 19:24 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 19:24 | |
@sonney2k | wiking, alright | 19:24 |
@sonney2k | wiking, fire it up | 19:24 |
wiking | ok going | 19:25 |
shogun-buildbot | build #92 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/92 | 19:25 |
shogun-buildbot | build #101 of deb3 - modular_interfaces started, including [7b3ee5ae63d2dcde25cc70e4074d75ecd11b0f98, 97763bbc134688dc412b6c735fa076d85e8448b9] | 19:25 |
shogun-buildbot | build #20 of cyg1 - libshogun started, including [7b3ee5ae63d2dcde25cc70e4074d75ecd11b0f98, 97763bbc134688dc412b6c735fa076d85e8448b9, 69d42744dff01bf9208151d829fa7e204926227d, 643ee74dc7f43c93dc27c7be62c7349079dc17b5, 28c06ce21772e79529bf3caa39f8b484bae0dcff] | 19:25 |
@sonney2k | wiking, when it works it should be listed here http://shogun-toolbox.org/buildbot/buildslaves | 19:25 |
wiking | yep | 19:25 |
wiking | but something went wrong ;) | 19:25 |
@sonney2k | wiking, authentification problem | 19:26 |
@sonney2k | did you use shogun-freebsd1 as name? | 19:26 |
@sonney2k | wiking, when you vim buildbot.tac | 19:26 |
@sonney2k | look at slavename | 19:26 |
@sonney2k | it should be 'shogun-freebsd1' | 19:26 |
wiking | oh no | 19:26 |
wiking | it's just shogun-freebsd | 19:27 |
wiking | but i'l change | 19:27 |
wiking | ok | 19:27 |
@sonney2k | yeah change it and buildslave start again | 19:27 |
@sonney2k | http://shogun-toolbox.org/buildbot/buildslaves | 19:27 |
@sonney2k | there we go | 19:27 |
@sonney2k | shogun-buildbot, force build "bsd1 - libshogun" initial | 19:28 |
shogun-buildbot | build #0 of bsd1 - libshogun started, including [] | 19:28 |
shogun-buildbot | build #0 forced | 19:28 |
shogun-buildbot | I'll give a shout when the build finishes | 19:28 |
@sonney2k | wiking, does it do anything? | 19:28 |
@sonney2k | wiking, it should checkout the sources via git | 19:29 |
@sonney2k | ~200 MB | 19:29 |
@sonney2k | that will take a while on the first test... | 19:30 |
wiking | yep | 19:31 |
wiking | it's checking out now the submodule | 19:31 |
blackburn | sonney2k: farmer | 19:32 |
shogun-buildbot | build #0 of bsd1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/0 | 19:32 |
wiking | eh | 19:33 |
@sonney2k | blackburn, ? | 19:33 |
blackburn | sonney2k: you're growing a farm of bots :) | 19:35 |
shogun-buildbot | build #20 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/20 | 19:36 |
wiking | sonney2k: ok this won't work | 19:36 |
wiking | cd src && ./configure --cc=gcc46 --cxx=c++46 --interfaces= --prefix=`pwd`/build' | 19:37 |
wiking | shouldn't it be cd source/src & ./configure --cc=gcc46 --cxx=c++46 --interfaces= --prefix=../../build | 19:37 |
wiking | ? | 19:37 |
wiking | mm | 19:38 |
wiking | actually i dont know | 19:38 |
wiking | i have a bsd1_-_libshogun dir under which there are 2 dirs: build and source | 19:38 |
@sonney2k | wiking, do you have /bin/bash ? | 19:39 |
wiking | but well ... it tried under /usr/home/wiking/shogun-buildbot/bsd1_-_libshogun/build | 19:39 |
wiking | sonney2k: nono, freebsd doesn't have that | 19:39 |
wiking | :) | 19:39 |
wiking | by default | 19:39 |
wiking | i have /usr/local/bin/bash | 19:39 |
@sonney2k | ok that is the reason then | 19:40 |
wiking | but try rather /usr/bin/env bash | 19:40 |
wiking | ;) | 19:40 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 19:40 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 19:40 | |
wiking | but i guess a /bin/sh would be sufficient as well | 19:40 |
@sonney2k | wiking, please restart the slave | 19:41 |
@sonney2k | buildslave stop | 19:41 |
@sonney2k | buildslave start | 19:41 |
wiking | ok | 19:41 |
blackburn | sonney2k: do you mind to output .o files to shogun/build dir? | 19:41 |
@sonney2k | blackburn, ugh not so easy | 19:41 |
wiking | sonney2k: restarted... | 19:41 |
blackburn | sonney2k: do you mind me to try? | 19:41 |
@sonney2k | blackburn, we have to change the makefile and everything | 19:41 |
@sonney2k | shogun-buildbot, force build "bsd1 - libshogun" initial | 19:42 |
shogun-buildbot | build #1 of bsd1 - libshogun started, including [] | 19:42 |
shogun-buildbot | build #1 forced | 19:42 |
shogun-buildbot | I'll give a shout when the build finishes | 19:42 |
@sonney2k | blackburn, if I was you I would rather do the doc or other important stuff now | 19:42 |
blackburn | heh | 19:42 |
@sonney2k | blackburn, that is some huge undertaking | 19:42 |
blackburn | is it really so hard? | 19:42 |
wiking | ok now it's running afaik | 19:42 |
@sonney2k | wiking, yeah looks good | 19:43 |
CIA-18 | shogun: Heiko Strathmann master * re1db78e / (2 files): removed compiler warning - http://git.io/MRu-zg | 19:43 |
CIA-18 | shogun: Heiko Strathmann master * r3e19962 / src/shogun/statistics/HSIC.cpp : fixes, gamma test should work now - http://git.io/cjtv4Q | 19:43 |
CIA-18 | shogun: Heiko Strathmann master * r76ba56d / src/shogun/statistics/HSIC.h : comment update - http://git.io/17Vy4A | 19:43 |
CIA-18 | shogun: Heiko Strathmann master * r564e0bb / examples/undocumented/libshogun/statistics_hsic.cpp : added test for HSIC gamma based test - http://git.io/C6CzCA | 19:43 |
CIA-18 | shogun: Heiko Strathmann master * rf842ad6 / (5 files in 2 dirs): Merge pull request #631 from karlnapf/master - http://git.io/KsKfGQ | 19:43 |
heiko | blackburn, did you already found the error? | 19:43 |
@sonney2k | wiking, now please continue with the other pull requests :D | 19:43 |
blackburn | heiko: not really | 19:44 |
blackburn | I stopped with getting proper results for now | 19:44 |
blackburn | :D | 19:44 |
wiking | sonney2k: alrighty | 19:44 |
blackburn | heiko: print_result of CVEvaluation prints last evaluation result I am afraid | 19:44 |
@sonney2k | wiking, at some point we could add more to the freebsd buildbot, like static and dynamic interfaces | 19:44 |
@sonney2k | maybe not all of them (you probably don't have csharp/java or?) | 19:45 |
wiking | sonney2k: java i do | 19:45 |
wiking | but not mono | 19:45 |
heiko | really? | 19:45 |
heiko | so not the best result | 19:45 |
heiko | Ill check the code | 19:45 |
shogun-buildbot | build #97 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/97 blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com> | 19:46 |
shogun-buildbot | build #93 of deb1 - libshogun started, including [e1db78e59c60dc0111d049dca777eb19762d7e03, 3e19962f1031075d3c4404d50d0b7a4ad36f465f, 76ba56d9a2dc191cfbbcdb6534c62b1912c803e6, 564e0bbab9c410e51319c054b15ead0087e58793, f842ad6f54c515880bf9e4c3b6ac387dc0d1dcb4] | 19:46 |
blackburn | heiko: | 19:46 |
blackburn | first.py:33: RuntimeWarning: [WARN] FeatureBlockLogisticRegression does not support locking. Autolocking is skipped. Set autolock flag to false to get rid of warning. | 19:46 |
blackburn | cross_validation.evaluate().print_result() | 19:46 |
blackburn | 0.500000 | 19:46 |
blackburn | it does print 0.5 | 19:46 |
blackburn | and I get 0.78 actually | 19:46 |
heiko | oh blackburn, looking at your code thats no surprise | 19:46 |
heiko | you apply the best parameters to the model *after* you print | 19:47 |
blackburn | what did I do wrong? | 19:47 |
heiko | change the two lines | 19:47 |
heiko | cross_validation.evaluate().print_result()best_parameters.apply_to_machine(lr) | 19:47 |
blackburn | ohhh I see | 19:47 |
shogun-buildbot | build #1 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/1 | 19:48 |
blackburn | heiko: my bad, thanks | 19:50 |
heiko | no worries :) | 19:50 |
shogun-buildbot | build #93 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/93 | 19:50 |
shogun-buildbot | build #103 of deb3 - modular_interfaces started, including [e1db78e59c60dc0111d049dca777eb19762d7e03] | 19:50 |
shogun-buildbot | build #2 of bsd1 - libshogun started, including [e1db78e59c60dc0111d049dca777eb19762d7e03, 3e19962f1031075d3c4404d50d0b7a4ad36f465f, 76ba56d9a2dc191cfbbcdb6534c62b1912c803e6, 564e0bbab9c410e51319c054b15ead0087e58793, f842ad6f54c515880bf9e4c3b6ac387dc0d1dcb4] | 19:50 |
shogun-buildbot | build #21 of cyg1 - libshogun started, including [e1db78e59c60dc0111d049dca777eb19762d7e03, 3e19962f1031075d3c4404d50d0b7a4ad36f465f, 76ba56d9a2dc191cfbbcdb6534c62b1912c803e6, 564e0bbab9c410e51319c054b15ead0087e58793, f842ad6f54c515880bf9e4c3b6ac387dc0d1dcb4] | 19:50 |
blackburn | heiko: it must be funny but it is the first time I use your framework :D | 19:50 |
heiko | blackburn, really? well, hope it works :D | 19:51 |
shogun-buildbot | build #21 of cyg1 - libshogun is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/21 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 19:51 |
heiko | Its good when people use it because errors are found | 19:51 |
@sonney2k | that's crazy | 19:51 |
blackburn | heiko: with my road sign shit I just get proper C and nothing more | 19:51 |
@sonney2k | https://www.youtube.com/watch?v=tvxJs_4m0ZE | 19:51 |
heiko | its better that you find them instead poeple that try to use shogun | 19:51 |
blackburn | sonney2k: back to youtube? | 19:51 |
@sonney2k | wiking, can you install glpk on that machine | 19:53 |
heiko | errrr | 19:53 |
heiko | sonney2k, have a look at the last error | 19:53 |
heiko | ah | 19:54 |
blackburn | heiko: I will use it in my next dirty deals for sure | 19:54 |
heiko | thats windows | 19:54 |
heiko | nice | 19:54 |
heiko | blackburn, yeah :D | 19:54 |
@sonney2k | wiking, quite a bit of stuff is not found - like libjson, superlu, arpack, nlopt, glpk, snappy, bzip2, lzma | 19:55 |
shogun-buildbot | build #2 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/2 | 19:56 |
CIA-18 | shogun: Soeren Sonnenburg master * r7cfd299 / src/shogun/io/AsciiFile.cpp : fix unsigned / signed comparison warning - http://git.io/kogVNg | 19:59 |
wiking | sonney2k: let's see | 19:59 |
@sonney2k | wiking, if it is easy please install these things | 19:59 |
wiking | sonney2k: that'll require some ./configure changes imo | 20:00 |
wiking | sonney2k: e.g. bzip2 is installed ;) | 20:01 |
wiking | it's part of the system | 20:01 |
@sonney2k | wiking, not the executable | 20:01 |
@sonney2k | the bzip2 headers | 20:01 |
wiking | as well | 20:01 |
@sonney2k | shogun has support for compressing / uncompressing such stuff | 20:01 |
wiking | btw arpack96 is good enough? | 20:02 |
wiking | or arpack-ng would be better? | 20:02 |
@sonney2k | wiking, then have a look at configure.log in that machine | 20:02 |
blackburn | arpack-ng? | 20:02 |
blackburn | wiking: 96 should be ok | 20:02 |
@sonney2k | wiking, it is called arpack 3.1.1 here | 20:02 |
wiking | blackburn: ARPACK-NG is a collection of Fortran77 subroutines designed to solve | 20:02 |
wiking | large-scale eigenvalue problems. It is a fork of the Rice University | 20:02 |
wiking | ARPACK, and is jointly-maintained by Debian, Octave, and Scilab. | 20:02 |
shogun-buildbot | build #102 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/102 blamelist: Soeren Sonnenburg <sonne@debian.org>, Viktor Gal <viktor.gal@maeth.com> | 20:02 |
shogun-buildbot | build #104 of deb3 - modular_interfaces started, including [3e19962f1031075d3c4404d50d0b7a4ad36f465f, 76ba56d9a2dc191cfbbcdb6534c62b1912c803e6, 564e0bbab9c410e51319c054b15ead0087e58793, f842ad6f54c515880bf9e4c3b6ac387dc0d1dcb4] | 20:02 |
wiking | sonney2k: ok then that's arpack-ng ;) | 20:03 |
blackburn | wiking: ok | 20:03 |
blackburn | I should get rid of it | 20:03 |
blackburn | :D | 20:03 |
@sonney2k | blackburn, yoyo! | 20:03 |
blackburn | sonney2k: what? | 20:03 |
@sonney2k | shogun-buildbot, force build "cyg1 - libshogun" wtf? | 20:05 |
shogun-buildbot | build #22 of cyg1 - libshogun started, including [] | 20:05 |
shogun-buildbot | build #22 forced | 20:05 |
shogun-buildbot | I'll give a shout when the build finishes | 20:05 |
@sonney2k | heiko, ^ lets retry - my guess it that it was somehow still running while we were killing / restarting the buildmaster | 20:06 |
heiko | sonney2k, yeah, I though I had broken something again :) | 20:06 |
@sonney2k | heiko, can you do anything about the current crashers? | 20:08 |
heiko | sonney2k, blackburn, has something changed in context of threads and kernel machines? | 20:08 |
@sonney2k | the real ones I mean | 20:08 |
heiko | Currently investigating | 20:08 |
@sonney2k | heiko, why? waht? | 20:08 |
heiko | I have the suspiction that subsets are used /set/unset from multiple threads | 20:08 |
heiko | which *will* crash | 20:08 |
heiko | and might explain the error where assertion fails that there is a different number of labels | 20:09 |
heiko | any changes in that recently? | 20:09 |
@sonney2k | blackburn,^? | 20:10 |
blackburn | hm | 20:10 |
@sonney2k | I didn't change anything there | 20:10 |
blackburn | what did I change is zeroing of result vector | 20:10 |
blackburn | in case of batch evaluation | 20:11 |
@sonney2k | yeah that was OK I think | 20:11 |
heiko | ok Ill check a bi tmore | 20:11 |
@sonney2k | blackburn, I don't understand though what we had there initially | 20:11 |
blackburn | sonney2k: we had no sgvector before here I guess | 20:11 |
@sonney2k | blackburn, yeah but then we should have had the same error... | 20:12 |
blackburn | sonney2k: if there was calloc before - no | 20:12 |
@sonney2k | float64_t* output=SG_MALLOC(float64_t, num_vectors); | 20:13 |
@sonney2k | memset(output, 0, sizeof(float64_t)*num_vectors); | 20:13 |
@sonney2k | that was there before | 20:13 |
blackburn | sonney2k: then the bug we introduced | 20:13 |
@sonney2k | I actually | 20:14 |
@sonney2k | for some reason I had in mind that sgvector uses CALLOC | 20:15 |
@sonney2k | well | 20:15 |
shogun-buildbot | build #22 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/22 | 20:17 |
@sonney2k | heiko, well try to reproduce it with #threads = 1 | 20:18 |
shogun-buildbot | build #103 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/103 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 20:20 |
shogun-buildbot | build #98 of deb2 - static_interfaces started, including [e1db78e59c60dc0111d049dca777eb19762d7e03, 3e19962f1031075d3c4404d50d0b7a4ad36f465f, 76ba56d9a2dc191cfbbcdb6534c62b1912c803e6, 564e0bbab9c410e51319c054b15ead0087e58793, f842ad6f54c515880bf9e4c3b6ac387dc0d1dcb4] | 20:20 |
@sonney2k | evaluation_cross_validation_locked_comparison is the only one crashing now | 20:20 |
heiko | yeh | 20:20 |
heiko | sonney2k, how to manually set num threads | 20:21 |
blackburn | heiko: .parallel.set_num_threads(1) | 20:23 |
heiko | thanks | 20:24 |
heiko | ok its not thread related | 20:25 |
shogun-buildbot | build #104 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/104 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 20:25 |
shogun-buildbot | build #94 of deb1 - libshogun started, including [7cfd2994914132c032d2a729d9a3185657cb57d3] | 20:25 |
shogun-buildbot | build #98 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/98 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 20:25 |
shogun-buildbot | build #94 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/94 | 20:25 |
shogun-buildbot | build #99 of deb2 - static_interfaces started, including [7cfd2994914132c032d2a729d9a3185657cb57d3] | 20:25 |
shogun-buildbot | build #105 of deb3 - modular_interfaces started, including [7cfd2994914132c032d2a729d9a3185657cb57d3] | 20:25 |
shogun-buildbot | build #3 of bsd1 - libshogun started, including [7cfd2994914132c032d2a729d9a3185657cb57d3] | 20:25 |
shogun-buildbot | build #23 of cyg1 - libshogun started, including [7cfd2994914132c032d2a729d9a3185657cb57d3] | 20:25 |
shogun-buildbot | build #99 of deb2 - static_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/99 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:27 |
blackburn | http://www.johndcook.com/blog/2012/07/07/wrong-and-unnecessary/ | 20:28 |
blackburn | sonney2k: that bot thing becomes too verbose I think | 20:28 |
shogun-buildbot | build #3 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/3 | 20:32 |
shogun-buildbot | build #23 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/23 | 20:36 |
heiko | blackburn, it was you! | 20:40 |
heiko | I knew it | 20:41 |
heiko | you are evil! ;) | 20:41 |
blackburn | heiko: what did I do? | 20:41 |
heiko | why did you comment out lines 268 ... from KernelMachine? | 20:41 |
heiko | that breaks apply(CFeatures*) | 20:41 |
blackburn | damn | 20:41 |
heiko | and since this example of mine is the only test where this is covered it fails | 20:41 |
blackburn | it was mistake of mine cause I did that only for test | 20:42 |
heiko | its alright | 20:42 |
heiko | Ill put it in again | 20:42 |
blackburn | I didn't mean to break your heart | 20:42 |
blackburn | :D | 20:42 |
-!- blackburn [~blackburn@81.28.187.0] has quit [Quit: Leaving.] | 20:43 | |
heiko | lol ;) | 20:43 |
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun | 20:45 | |
CIA-18 | shogun: Heiko Strathmann master * r96f7942 / src/shogun/evaluation/CrossValidation.cpp : added some debug messages - http://git.io/69KaSw | 20:46 |
CIA-18 | shogun: Heiko Strathmann master * r1bc6e55 / src/shogun/evaluation/ContingencyTableEvaluation.cpp : added a proper error message if number of labels are different - http://git.io/yhgktg | 20:46 |
CIA-18 | shogun: Heiko Strathmann master * rc593061 / (3 files in 2 dirs): Merge pull request #632 from karlnapf/master - http://git.io/sQBk3Q | 20:46 |
CIA-18 | shogun: Heiko Strathmann master * r94a6183 / src/shogun/machine/KernelMachine.cpp : removed /* */ to make locked evaluation example work again - http://git.io/IJSdXQ | 20:46 |
shogun-buildbot | build #95 of deb1 - libshogun started, including [1bc6e550a01be7d531c422edf108dd7cb8236faa] | 20:50 |
shogun-buildbot | build #95 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/95 | 20:51 |
shogun-buildbot | build #96 of deb1 - libshogun started, including [96f79427c669926dca226b44fba02559a67fac5b, 94a618324be3c5f4d6791372f849904a2e40edfa, c593061648fcc2f3b8908509074d0e8eb702211d] | 20:51 |
CIA-18 | shogun: Sergey Lisitsyn master * r3da93b3 / src/shogun/lib/IndexBlockTree.cpp : Removed root feature from G/ind - http://git.io/ZqwmwQ | 20:51 |
CIA-18 | shogun: Sergey Lisitsyn master * rf39aa5c / src/shogun/lib/slep/slep_logistic.cpp : Commented out gradient and objective value debug prints - http://git.io/mfPotQ | 20:51 |
shogun-buildbot | build #4 of bsd1 - libshogun started, including [1bc6e550a01be7d531c422edf108dd7cb8236faa] | 20:51 |
CIA-18 | shogun: Sergey Lisitsyn master * rc23a3c1 / src/shogun/modelselection/GridSearchModelSelection.cpp : Made evaluation direction verbose - http://git.io/T00Urw | 20:51 |
shogun-buildbot | build #24 of cyg1 - libshogun started, including [1bc6e550a01be7d531c422edf108dd7cb8236faa] | 20:51 |
blackburn | heiko: I can haz a few commits too | 20:51 |
shogun-buildbot | build #96 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/96 | 20:52 |
shogun-buildbot | build #106 of deb3 - modular_interfaces started, including [1bc6e550a01be7d531c422edf108dd7cb8236faa] | 20:52 |
heiko | blackburn, say what could be the reason that an example fails during make tests | 20:52 |
heiko | but works when I run it alone? | 20:52 |
blackburn | heiko: puzzles? | 20:52 |
heiko | whats that? | 20:52 |
blackburn | interesting | 20:52 |
blackburn | I mean you ask me for puzzle solve | 20:52 |
blackburn | :D | 20:52 |
heiko | hehe | 20:52 |
heiko | amazing | 20:52 |
heiko | runs with valgrind | 20:52 |
heiko | but not without | 20:52 |
blackburn | heiko: I can't mind any reason | 20:52 |
blackburn | ah no | 20:53 |
blackburn | that is not so interesting | 20:53 |
blackburn | it happens often | 20:53 |
heiko | and no leak | 20:53 |
heiko | why is that? | 20:53 |
heiko | actually many many libshogun examples are failing locally here | 20:53 |
heiko | with that problem | 20:53 |
heiko | why is it? | 20:53 |
blackburn | with which problem | 20:53 |
blackburn | ? | 20:53 |
heiko | /home/heiko/Desktop/shogun/shogun/examples/undocumented/libshogun/modelselection_grid_search_string_kernel.cpp | 20:54 |
heiko | for example | 20:54 |
heiko | yeah | 20:54 |
heiko | valgrind makes it run | 20:54 |
heiko | but it doesnt without | 20:54 |
blackburn | heiko: segmentation fault or what? | 20:54 |
heiko | and I get strange CMap assertion errors | 20:54 |
heiko | WTF | 20:55 |
heiko | thats my commetn | 20:55 |
shogun-buildbot | build #105 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/105 blamelist: Soeren Sonnenburg <sonne@debian.org> | 20:55 |
shogun-buildbot | build #100 of deb2 - static_interfaces started, including [1bc6e550a01be7d531c422edf108dd7cb8236faa, 96f79427c669926dca226b44fba02559a67fac5b, 94a618324be3c5f4d6791372f849904a2e40edfa, c593061648fcc2f3b8908509074d0e8eb702211d] | 20:55 |
shogun-buildbot | build #4 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/4 | 20:57 |
shogun-buildbot | build #5 of bsd1 - libshogun started, including [96f79427c669926dca226b44fba02559a67fac5b, 94a618324be3c5f4d6791372f849904a2e40edfa, c593061648fcc2f3b8908509074d0e8eb702211d] | 20:57 |
heiko | sonney2k, any idea? an example runs fine with valgrind(no leaks, no errors), but assertion fails without valgrind | 20:57 |
heiko | sonney2k, also, many examples fail here while they dont on the buildbot | 21:00 |
heiko | I have 64bit, does anybody else have? | 21:00 |
heiko | blackburn? ^ | 21:00 |
blackburn | heiko: 64 here | 21:00 |
heiko | and make tests works? | 21:00 |
blackburn | let me try | 21:01 |
shogun-buildbot | build #5 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/5 | 21:02 |
shogun-buildbot | build #24 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/24 | 21:03 |
shogun-buildbot | build #25 of cyg1 - libshogun started, including [96f79427c669926dca226b44fba02559a67fac5b, 94a618324be3c5f4d6791372f849904a2e40edfa, c593061648fcc2f3b8908509074d0e8eb702211d] | 21:03 |
shogun-buildbot | build #100 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/100 | 21:04 |
shogun-buildbot | build #107 of deb3 - modular_interfaces started, including [96f79427c669926dca226b44fba02559a67fac5b, 94a618324be3c5f4d6791372f849904a2e40edfa, c593061648fcc2f3b8908509074d0e8eb702211d] | 21:04 |
heiko | blackburn, sonney2k, what do you think of gnu-intend ? | 21:04 |
blackburn | heiko: never tried but I don't mind to reformat everything | 21:04 |
heiko | I think that would be an idea, esp. doing it automatically | 21:05 |
blackburn | heiko: all examples passed | 21:05 |
heiko | strange! | 21:05 |
heiko | here, no | 21:05 |
heiko | do you have trace-mallocs enabled? | 21:06 |
heiko | blackburn, ^ | 21:07 |
blackburn | no | 21:07 |
heiko | mmh | 21:07 |
heiko | that might be the reason | 21:07 |
heiko | Ill disable it | 21:07 |
blackburn | hm CMap? | 21:09 |
blackburn | I guess it is | 21:09 |
heiko | yeh | 21:10 |
heiko | yeah its CMap | 21:11 |
heiko | runs when I deactivate trace-mallocs | 21:11 |
heiko | but the valgrind things is even more strange | 21:11 |
heiko | blackburn, also your example all fail when trace-mallocs is on | 21:12 |
heiko | do you use CMap? | 21:12 |
blackburn | no | 21:12 |
blackburn | trace-mallocs does | 21:12 |
blackburn | heiko: which mine? | 21:12 |
heiko | in the embeddings | 21:12 |
heiko | mmh | 21:13 |
heiko | Ill recompile and tell you | 21:13 |
blackburn | argh | 21:13 |
blackburn | shouldn't be actually | 21:13 |
shogun-buildbot | build #25 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/25 | 21:14 |
shogun-buildbot | build #106 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/106 blamelist: Heiko Strathmann <heiko.strathmann@gmail.com> | 21:14 |
shogun-buildbot | build #97 of deb1 - libshogun started, including [3da93b3ce0d4cec8552087c5c8b91c9426cfc06a, f39aa5cb80d23744188634d098881456de6763f4, c23a3c167b150ab6bd0922665e72b5d16fb4992f] | 21:14 |
shogun-buildbot | build #97 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/97 | 21:15 |
shogun-buildbot | build #108 of deb3 - modular_interfaces started, including [3da93b3ce0d4cec8552087c5c8b91c9426cfc06a, f39aa5cb80d23744188634d098881456de6763f4] | 21:15 |
shogun-buildbot | build #26 of cyg1 - libshogun started, including [3da93b3ce0d4cec8552087c5c8b91c9426cfc06a, f39aa5cb80d23744188634d098881456de6763f4, c23a3c167b150ab6bd0922665e72b5d16fb4992f] | 21:15 |
shogun-buildbot | build #6 of bsd1 - libshogun started, including [3da93b3ce0d4cec8552087c5c8b91c9426cfc06a, f39aa5cb80d23744188634d098881456de6763f4, c23a3c167b150ab6bd0922665e72b5d16fb4992f] | 21:15 |
heiko | blackburn, | 21:18 |
heiko | ./converter_localtangentspacealignment | 21:18 |
heiko | ./converter_hessianlocallylinearembedding | 21:19 |
heiko | ./converter_kernellocaltangentspacealignment | 21:19 |
heiko | ./converter_linearlocaltangentspacealignment | 21:19 |
heiko | try make tests with --enable-trace_mallocs | 21:19 |
blackburn | will do | 21:19 |
shogun-buildbot | build #6 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/6 | 21:21 |
shogun-buildbot | build #26 of cyg1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/26 | 21:26 |
heiko | will go now, see you tomorrow! bye | 21:29 |
-!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has left #shogun [] | 21:29 | |
@sonney2k | blackburn, so do things fail for you when trace mallocs is on? | 21:31 |
blackburn | running | 21:32 |
blackburn | sonney2k: yeah they fail here | 21:33 |
@sonney2k | well ok lets see then | 21:34 |
@sonney2k | blackburn, about KernelMachine.cpp | 21:36 |
@sonney2k | line 263 | 21:36 |
@sonney2k | I think it should do init() even if kernel is COMBINED... | 21:38 |
@sonney2k | let me change that | 21:39 |
shogun-buildbot | build #107 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/107 | 21:39 |
shogun-buildbot | build #101 of deb2 - static_interfaces started, including [3da93b3ce0d4cec8552087c5c8b91c9426cfc06a, f39aa5cb80d23744188634d098881456de6763f4, c23a3c167b150ab6bd0922665e72b5d16fb4992f] | 21:39 |
CIA-18 | shogun: Soeren Sonnenburg master * r3702954 / src/shogun/machine/KernelMachine.cpp : init(lhs,data) even for combined kernel in apply() - http://git.io/-ocBHw | 21:41 |
@sonney2k | blackburn, about the bots - what do you suggest? | 21:42 |
blackburn | sonney2k: no idea, what can I suggest? | 21:42 |
shogun-buildbot | build #108 of deb3 - modular_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/108 | 21:44 |
shogun-buildbot | build #109 of deb3 - modular_interfaces started, including [c23a3c167b150ab6bd0922665e72b5d16fb4992f] | 21:44 |
shogun-buildbot | build #101 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/101 | 21:47 |
shogun-buildbot | build #98 of deb1 - libshogun started, including [3702954a9ef3a0418a7ce4103e2551206191d1fd] | 21:47 |
@sonney2k | blackburn, well we can make the notify us on basically anything or not, e.g. start of the build end of the build | 21:49 |
@sonney2k | only when things fail | 21:49 |
@sonney2k | or when things go from fail -> success | 21:49 |
@sonney2k | etc | 21:49 |
shogun-buildbot | build #98 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/98 | 21:52 |
shogun-buildbot | build #102 of deb2 - static_interfaces started, including [3702954a9ef3a0418a7ce4103e2551206191d1fd] | 21:52 |
shogun-buildbot | build #7 of bsd1 - libshogun started, including [3702954a9ef3a0418a7ce4103e2551206191d1fd] | 21:52 |
shogun-buildbot | build #27 of cyg1 - libshogun started, including [3702954a9ef3a0418a7ce4103e2551206191d1fd] | 21:52 |
blackburn | sonney2k: I'd like to see notifications only when state changes | 21:54 |
@sonney2k | blackburn, I think we should also get notified as long as things are still broken | 21:55 |
blackburn | yeah may be | 21:55 |
@sonney2k | otherwise people start to relax even though stuff is broken | 21:56 |
@sonney2k | blackburn, we can have different notifications in IRC and via email btw | 21:56 |
@sonney2k | blackburn, so for IRC you want success -> failure, failure -> success, failure only right? | 21:57 |
@sonney2k | not started, finished, success, exception? | 21:58 |
blackburn | sonney2k: yeah | 21:58 |
blackburn | success -> success is not of interest | 21:58 |
blackburn | exception and not started is rather for you :) | 21:58 |
shogun-buildbot | build #7 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/7 | 21:58 |
@sonney2k | blackburn, nope | 21:58 |
@sonney2k | no one notices if sth is not starting... | 21:58 |
shogun-buildbot | build #102 of deb2 - static_interfaces is complete: Failure [failed test octave_static] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/102 blamelist: Soeren Sonnenburg <sonne@debian.org> | 21:59 |
shogun-buildbot | build #110 of deb3 - modular_interfaces started, including [3702954a9ef3a0418a7ce4103e2551206191d1fd] | 21:59 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection] | 22:01 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 22:01 | |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun | 22:04 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:09 | |
blackburn | sonney2k: do you mind to change to openmp next week? | 22:15 |
blackburn | n4nd0: can you invest some time? | 22:16 |
blackburn | next week, again | 22:16 |
@sonney2k | blackburn, if you can get rid of pthreads and make sure everything is at same speed sure | 22:16 |
blackburn | sonney2k: no way to make it in one commit but I can start | 22:16 |
n4nd0 | blackburn: tell me | 22:17 |
blackburn | n4nd0: I'd ask you to help me with openmping | 22:18 |
n4nd0 | blackburn: the only thing I know is that it is something related to multithreading :) | 22:18 |
blackburn | quite easy | 22:18 |
n4nd0 | blackburn: but it would be cool to learn about it! | 22:18 |
blackburn | you just declare pragma omp with some parameters | 22:18 |
n4nd0 | blackburn: mmm ok, so tell me more in particular what needs to be done and so | 22:19 |
blackburn | n4nd0: we have a few classes using pthreads | 22:19 |
blackburn | all one need is to get it serial first | 22:20 |
blackburn | and then declare OMP pragma before 'for' statement | 22:20 |
blackburn | sonney2k: so ok to start? | 22:20 |
n4nd0 | blackburn: has something been ported already so I can take a look? | 22:21 |
blackburn | n4nd0: yeah it was somewhere in my branches | 22:21 |
blackburn | I hope I didn't lost it | 22:21 |
blackburn | yeaaah I didn't! | 22:22 |
blackburn | would be hard to merge probably now | 22:22 |
blackburn | let me try | 22:22 |
blackburn | n4nd0: are you busy right now? | 22:27 |
@sonney2k | blackburn, I prefer one commit for pthread -> openmp - it doesn't really help us to cause new trouble/dependencies | 22:28 |
blackburn | sonney2k: are you kidding? | 22:28 |
@sonney2k | and I really don't know if we can change *everything* to be as fast as it was with pthreads | 22:29 |
@sonney2k | blackburn, no | 22:29 |
@sonney2k | http://www.legoturingmachine.org/ | 22:29 |
blackburn | it helps us | 22:29 |
blackburn | lesser code is better | 22:29 |
@sonney2k | blackburn, of course but the whole purpose of parallelizing things is that it gets faster | 22:29 |
blackburn | sonney2k: openmp *can't* be slower | 22:29 |
@sonney2k | and I failed back then with openmp | 22:30 |
blackburn | with proper options | 22:30 |
@sonney2k | haha | 22:30 |
@sonney2k | create a branch for that and push it / and keep it regularly in sync | 22:30 |
blackburn | sonney2k: I already tested CDistance and openmp is even slightly faster | 22:30 |
@sonney2k | but one push to master for that | 22:30 |
blackburn | hmmm ok I don't mind separate branch | 22:30 |
@sonney2k | that is no guarantee that it works on the other stuff too | 22:31 |
n4nd0 | blackburn: yeah sorry, I am trying to buy plane tickets to Spain | 22:31 |
blackburn | heh | 22:31 |
n4nd0 | blackburn: I will be back here in a few minutes | 22:31 |
blackburn | going to spain? | 22:31 |
n4nd0 | blackburn: some days in August yeah | 22:31 |
@sonney2k | blackburn, I recall doing the ocas parallelization with openmp and it was just slow | 22:34 |
@sonney2k | all CPUs had 100% load | 22:34 |
@sonney2k | but no speedup | 22:34 |
blackburn | sonney2k: did you use nowait, dynamic schedule? | 22:34 |
blackburn | with different scheduling routines I've got different results | 22:35 |
@sonney2k | no idea - did that about 7 years ago... | 22:35 |
@sonney2k | blackburn, but I really think that now that n4nd0, pluskid etc have some .pdf - documentation is much more important right now | 22:45 |
blackburn | wellllllll | 22:45 |
blackburn | try to push them then :D | 22:45 |
@sonney2k | blackburn, they did the work already | 22:46 |
@sonney2k | so we are waiting for your shogun-doc git repo to merge the stuff | 22:46 |
blackburn | ok I'll come up with template and start to merge then | 22:47 |
@sonney2k | thanks | 22:48 |
blackburn | sonney2k: but I remember you didn't like the book idea | 22:49 |
blackburn | what has changed? | 22:49 |
@sonney2k | what? | 22:49 |
@sonney2k | I said that for a book it needs some person pushing it for probably 2 years | 22:50 |
blackburn | who is that person then? | 22:50 |
@sonney2k | no idea? you? | 22:50 |
blackburn | :D | 22:50 |
@sonney2k | but this here is about to have *some* documentation that is more user friendly | 22:51 |
@sonney2k | blackburn, if you manage to get the others into the boat then it might be possible to keep it up-to-date | 22:52 |
@sonney2k | otherwise what we have (documentation in the code) is the only thing that will remotely be up-to-date | 22:52 |
blackburn | sonney2k: should tutorial be submodule of shogun? | 23:06 |
shogun-buildbot | build #12 of nightly_all is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/12 | 23:11 |
n4nd0 | blackburn: all right, I am all yours now :) | 23:13 |
n4nd0 | I hate buying plane tickets by the way... | 23:13 |
blackburn | n4nd0: hmm | 23:13 |
blackburn | n4nd0: as sonney2k denied pushing to master I should think how should we proceed | 23:13 |
n4nd0 | blackburn: ok | 23:13 |
n4nd0 | blackburn: we can have a branch where we work | 23:14 |
n4nd0 | blackburn: I can checkout yours maybe | 23:14 |
blackburn | n4nd0: you may help me with benchmark | 23:14 |
n4nd0 | blackburn: sure | 23:14 |
n4nd0 | blackburn: what plan do you have in mind? | 23:15 |
blackburn | n4nd0: I need some snippet to test performance of Kernel and Distance now | 23:16 |
n4nd0 | blackburn: any kernel? to compare with and without multithreading or? | 23:17 |
blackburn | meanwhile I'll try to merge conflicts from my earlier work | 23:17 |
blackburn | n4nd0: just with 1,2,4 threads, Linear and Euclidean I think | 23:17 |
blackburn | but a few times | 23:17 |
blackburn | to make sure it is stat. significant :D | 23:18 |
shogun-buildbot | build #17 of nightly_default is complete: Failure [failed updatedoc] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/17 | 23:19 |
@sonney2k | blackburn, you can use a remote branch just not master | 23:19 |
@sonney2k | blackburn, re tutorial makes sense | 23:20 |
blackburn | sonney2k: yes sure | 23:20 |
shogun-buildbot | build #15 of nightly_none is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/15 | 23:25 |
shogun-buildbot | build #103 of deb2 - static_interfaces is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/103 | 23:27 |
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Ping timeout: 245 seconds] | 23:50 | |
--- Log closed Mon Jul 09 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!