Open in new window / Try shogun cloud
--- Log opened Tue Nov 20 00:00:17 2012
blackburnwiking: what do you mena?00:42
blackburntypename/class is not that relevant I believe00:42
wiking:>00:50
wikingmena i mena00:50
wiking:D00:50
blackburnwiking: so really, what did you want to change with typename?00:53
wikinglol man00:57
wikingi just wrote a how to code email00:57
wikingto shogun00:57
wiking;)00:57
blackburn:D00:58
blackburnsonney2k: yeah adding an inplace new helps01:02
-!- ptizoom [~christian@79-71-89-182.dynamic.dsl.as9105.com] has quit [Quit: Ex-Chat]03:20
-!- ptizoom [~christian@79-71-89-182.dynamic.dsl.as9105.com] has joined #shogun03:21
shogun-buildbotbuild #177 of nightly_default is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/17704:07
@sonney2kblackburn, ?05:22
@sonney2krecall that this is supposed to use new[]()05:22
-!- ptizoom [~christian@79-71-89-182.dynamic.dsl.as9105.com] has quit [Quit: Ex-Chat]07:09
-!- ptizoom [~christian@79-71-89-182.dynamic.dsl.as9105.com] has joined #shogun07:09
blackburnsonney2k: yeah07:41
-!- blackburn [~blackburn@188.168.3.49] has quit [Quit: Leaving.]08:41
-!- zxtx [~zv@24.146.226.58] has joined #shogun10:11
-!- ptizoom [~christian@79-71-89-182.dynamic.dsl.as9105.com] has quit [Quit: Ex-Chat]10:17
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun10:28
-!- zxtx [~zv@24.146.226.58] has quit [Ping timeout: 245 seconds]11:35
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed]16:00
-!- blackburn [~blackburn@37.61.180.47] has joined #shogun17:36
-!- zxtx [~zv@rrcs-24-213-167-69.nyc.biz.rr.com] has joined #shogun21:31
wikingwtf22:02
wikingare we cblas support list?22:02
blackburnhahah22:04
blackburnI don't mind to share22:04
blackburnwiking: that's actually cool to share useless things I know :D22:05
wiking:>22:10
blackburnwiking: I have a question for ya22:17
blackburnwiking: I have quite a few loose coupled source files in there: https://github.com/lisitsyn/tapkee/tree/master/tapkee22:18
blackburnlike routines folder22:18
blackburnbut here I use things I defined in tapkee_defines.hpp22:18
blackburnso going from the top to the bottom it works and compiles ok22:18
blackburnbut when I have say https://github.com/lisitsyn/tapkee/blob/master/tapkee/routines/matrix_operations.hpp22:19
blackburnwhere defines file is not included22:19
blackburnit is not really consistent22:19
blackburnbut including defines there again is confusing for me22:19
blackburnso there is a loop - it makes no troubles yet but it is ugly22:20
blackburnwhat do you think?22:20
wikingjust a sec22:21
-shogungit:#shogun- [shogun] sonney2k pushed 1 new commit to master: https://github.com/shogun-toolbox/shogun/commit/aa689b0541203ff51b2d02706f7b4be129f4bb2122:24
-shogungit:#shogun- shogun/master aa689b0 Soeren Sonnenburg: Template based malloc/new etc...22:24
blackburnsonney2k: did that help?22:24
blackburnand wtf did you glance over all code22:25
blackburn:D22:25
-shogungit:#shogun- [shogun] sonney2k pushed 1 new commit to master: https://github.com/shogun-toolbox/shogun/commit/63d4f32d8b1543b272444076ded9adda57dba98b22:27
-shogungit:#shogun- shogun/master 63d4f32 Soeren Sonnenburg: remove debug printf's22:27
@sonney2kblackburn, yes it works now22:27
@sonney2kblackburn, the issue is that from w/i memory.h I needed to include SGVector.h22:28
@sonney2kthat is causing severe pain22:28
blackburnI see22:30
@sonney2kso basically I had to move everything to .cpp files as much as possible22:30
@sonney2kbut we should do this anyways22:30
@sonney2kI will try if I can get forward decl. to work for SGVector in memory now...22:31
@sonney2khttp://seclists.org/fulldisclosure/2012/Nov/9422:37
@sonney2kuhoh22:37
@sonney2klets up there will be a fix soon ...22:37
@sonney2kbtw have you noticed - mloss.org mldata.org etc are down22:39
blackburnsonney2k: yes I did22:39
@sonney2kit's being hosted at TU22:40
@sonney2kbut I have absolutely no longer any influence on this...22:40
blackburntu sucks :D22:40
@sonney2kmight be that the department is moving to a new building22:41
shogun-buildbotbuild #542 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/54222:43
shogun-buildbotbuild #417 of cyg1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/cyg1%20-%20libshogun/builds/41722:43
-!- travis-ci [~travis-ci@ec2-107-21-192-51.compute-1.amazonaws.com] has joined #shogun22:47
travis-ci[travis-ci] it's Soeren Sonnenburg's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/328833622:47
-!- travis-ci [~travis-ci@ec2-107-21-192-51.compute-1.amazonaws.com] has left #shogun []22:47
shogun-buildbotbuild #81 of ubu1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/ubu1%20-%20libshogun/builds/8122:49
@sonney2kwiking, look some tiny dataset to play with ;) http://www.nytimes.com/2012/11/20/science/for-web-images-creating-new-technology-to-seek-and-find.html?smid=go-share22:58
-shogungit:#shogun- [shogun] sonney2k pushed 1 new commit to master: https://github.com/shogun-toolbox/shogun/commit/af103f3dd6176b993052d42c2f038e38ae74962f23:01
-shogungit:#shogun- shogun/master af103f3 Soeren Sonnenburg: simplify example23:01
wikingyeah imagenet23:01
wikingit's goodish23:01
blackburnsonney2k: we are going to have randomized pca in shougn once I merge things :)23:06
blackburnand randomized kpca23:07
blackburnand randomized everything but it is really relevant for dense methods like mds, pca and isomap23:07
@sonney2kwe need some huge data for that :)23:09
@sonney2kbtw, blackburn does ./evaluation_cross_validation_multiclass_mkl work on your machine?23:10
@sonney2kI mean will it give valgrind errors23:10
@sonney2k(with 'old' shogun')23:10
@sonney2kor wiking ^23:10
blackburnI should try23:10
blackburnold like before your commit?23:10
@sonney2kyes23:10
blackburngood catch that I didn't update it yet23:10
blackburnsegfaul23:11
@sonney2kblackburn, and valgrind?23:12
@sonney2kwhat does it say?23:12
@sonney2k(I assume that example was broken before...)23:12
blackburnhe said I am an asshole but let me check more careful23:12
blackburna lot of invalid read of size 823:12
@sonney2ktell me anything I don't know :P23:12
@sonney2ka lot of?23:12
blackburnyes, hundreds23:13
blackburnokay tens23:13
blackburn:D23:13
@sonney2k==14496== Invalid free() / delete / delete[] / realloc()23:13
@sonney2kI only have this23:13
blackburnthat's the last one23:13
@sonney2k==14496==  Address 0xb723e90 is 0 bytes inside a block of size 16 free'd23:13
@sonney2k==14496==    at 0x4C27D4E: free (vg_replace_malloc.c:427)23:13
@sonney2k==14496==    by 0x52107C9: shogun::CParseBuffer<double>::~CParseBuffer() (memory.h:95)23:13
@sonney2k==14496==    by 0x5210838: shogun::CParseBuffer<double>::~CParseBuffer() (ParseBuffer.h:273)23:13
@sonney2kblackburn, which shogun version do you have?23:14
blackburn1.1.023:14
@sonney2kgrrg23:14
blackburncommit 93dba0dd6313f4faf6e2436b653ce1caf16880e123:14
blackburnI have 'last night shogun'23:14
@sonney2kblackburn, that is b0rken then anyways23:14
blackburnshould ich update?23:15
@sonney2kcould you please go back to c9a8772920c07f9b4cd8e29f664918e097f38bab23:15
shogun-buildbotbuild #654 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/65423:15
@sonney2kstrange that the buildbot is happy23:15
blackburnit doesn't compile poor sonne23:17
blackburnsonney2k: what should I do?23:17
blackburn../shogun/lib/memory.h:41:24: error: expected initializer before '<' token23:17
@sonney2kblackburn, what? that is before my changes23:17
@sonney2kahh23:17
@sonney2kwell remove that function then23:18
blackburnI did what you told me master23:18
@sonney2kohh wait23:18
@sonney2kwrong version23:18
blackburnhah23:18
blackburn<type>23:18
blackburnjava like23:18
@sonney2k7866d19a4183ad68e64c0d58fc40992de9ff967e23:18
@sonney2kyeah java infiltrated me23:19
@sonney2kI can no longer think clearly23:19
blackburnI will wait for ArrayList in your code one day23:19
blackburnsonney2k: interview question - what is the difference between Vector and ArrayList?23:19
@sonney2kno clue23:20
@sonney2kwhat is Vector?23:20
blackburnthread-safe arraylist23:20
blackburnarray-list is not thread safe23:20
blackburnsonney2k: StringBuffer and StringBuilder may be?23:20
blackburn;)23:20
@sonney2kblackburn, so does 7866d19a4183ad68e64c0d58fc40992de9ff967e compile?23:20
@sonney2kblackburn, no idea what StringBuffer is23:20
@sonney2kbut let me gues23:20
@sonney2ks23:20
blackburnit is compiling23:20
@sonney2kthread safe sb :D23:20
blackburnsonney2k: bingo23:21
blackburnI have ????-6 machine so please wait23:21
@sonney2kthe number of libs / support classes in java is beyond control23:21
blackburnsonney2k: number of features in C++ is beyond human brain capabilities :D23:22
@sonney2kdefinitely23:22
@sonney2kblackburn, stop using templates :D23:22
blackburnno it gives me power I never had before23:22
blackburnI can fly!23:22
@sonney2kI realize that when I am coding in java my brain relaxes23:22
@sonney2kI thought dance...23:22
blackburnI realize I shit bricks when I see java code at my job I am leaving23:23
@sonney2kis there some obfuscated java code contest it could enter?23:24
blackburnsonney2k: no typical overbusinessed code23:25
blackburna lot of code23:25
blackburnit looks like I am serving iran nuclear programme helping with blas23:26
blackburn:D23:26
@sonney2kthey need all support they can get ;)23:26
@sonney2kblackburn, so any news - did it build ?23:26
blackburnyes23:26
blackburnbuilding examples23:26
blackburniran is calling sorry :D23:27
@sonney2kblackburn, well please check with valgrind too23:27
@sonney2k./evaluation_cross_validation_multiclass_mkl that is23:28
blackburn==26493== Invalid free() / delete / delete[] / realloc()23:28
blackburnone23:28
@sonney2kalright23:28
@sonney2kthat is the one I have now too23:28
@sonney2kthen feel free to go update to master23:28
@sonney2kI can then get sgvector forward decls to work23:29
@sonney2k:)23:29
@sonney2kso things will improve23:29
blackburncool23:30
@sonney2kso the downside now is that we will have ~100 malloc implementations23:32
@sonney2kso our .so will grow a bit23:32
@sonney2kbut the upside is that stuff like DynArray will work directly with SGVector*23:33
@sonney2kand we could even have some better trace mallocs for SGObjects23:33
@sonney2kI mean listing the objects even with refcount :)23:33
blackburnis our .so now 500 mb?23:34
blackburn:D23:34
blackburnsonney2k: do yuo understand why clang is so much better than gcc?23:35
blackburnI am confused with that thing - gcc has been here for years and it is slow, often fails with vectorization and has unreadable errors23:36
blackburnthat's a puzzle for me23:36
@sonney2k102M Nov 20 23:03 shogun/libshogun.so.13.023:37
@sonney2kblackburn, what size is yours?23:37
blackburnI will not tell you :D23:37
blackburnokay wait wait it is compiling23:37
@sonney2kI knew you have to compensate for sth23:37
@sonney2kdidn't know it is *that* :D23:38
@sonney2kblackburn, that is with optimizations enabled + debug symbols23:38
blackburnyeah I don't have 102M23:38
blackburnmeters23:38
blackburn:D23:38
blackburnoptimizations enabled + debug yeah23:39
blackburnbtw without debug it should be much less23:39
blackburntapkee is 20mb when debug is enabled23:39
blackburn406kb with no debug info23:39
@sonney2kwhen I strip it it is 20mb23:40
@sonney2kactually23:40
@sonney2k1323:40
@sonney2kMB23:40
blackburnmakes sense23:40
@sonney2kso shogun == 30 tapkees23:40
blackburnyeah 15 pair of slippers23:41
blackburnmakes sense again23:41
@sonney2khighheels, boots for whatever weather conditions you encounter23:41
@sonney2kand occassions23:42
blackburnsonney2k: no tapkee is slippers23:42
@sonney2kblackburn, so is it ~100MB here too?23:42
blackburncompiling..23:42
@sonney2kblackburn, sure just plain slippers23:42
blackburnsonney2k: I'll name next piece of code I write as borsch may be23:43
@sonney2kargh kid awake, n823:43
blackburn.a is 220 MB23:47
blackburn.so is 98.1 MB23:47
@sonney2kblackburn, and when you do strip -s on the .so?23:48
blackburn12.2 mb23:48
blackburnsonney2k: I didn't know about strip - what does it do?23:49
blackburnis it corrupted now?23:49
blackburnwhen I strip tapkee app it becomes 335 kb23:50
@sonney2kblackburn, no just debug symbols removed23:53
blackburncool23:53
@sonney2kalright then good enough23:54
@sonney2kbed time!23:54
blackburnsee you23:55
--- Log closed Wed Nov 21 00:00:17 2012