--- Log opened Mon Jan 30 00:00:19 2012 | ||
-!- blackburn [~qdrgsm@188.122.239.116] has quit [Quit: Leaving.] | 00:28 | |
-!- blackburn [~qdrgsm@188.122.239.116] has joined #shogun | 03:47 | |
CIA-11 | shogun: Sergey Lisitsyn master * r689b5ed / (6 files): Fixes and improvements of EDRT - http://git.io/TEIZ3Q | 03:48 |
---|---|---|
shogun-buildbot | build #465 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/465 | 04:14 |
CIA-11 | shogun: Sergey Lisitsyn master * ra253451 / (6 files in 2 dirs): CoverTree for EDRT - http://git.io/YEmcvA | 07:26 |
-!- blackburn [~qdrgsm@188.122.239.116] has quit [Quit: Leaving.] | 07:27 | |
-!- blackburn [~qdrgsm@188.122.239.116] has joined #shogun | 07:59 | |
CIA-11 | shogun: Soeren Sonnenburg master * r3da3d15 / src/shogun/ui/GUIConverter.cpp : fix compilation error when lapack is not available - http://git.io/cWOeAw | 09:12 |
CIA-11 | shogun: Soeren Sonnenburg master * race4ba9 / (3 files in 3 dirs): | 10:54 |
CIA-11 | shogun: include <clapack.h> directly | 10:54 |
CIA-11 | shogun: we cannot include <atlas/clapack.h> as some distriutions ship clapack | 10:54 |
CIA-11 | shogun: under /usr/include - http://git.io/4WtslQ | 10:54 |
CIA-11 | shogun: Soeren Sonnenburg master * r66ebf04 / src/shogun/features/BinnedDotFeatures.cpp : fix normalization in binned dot features' dot() function - http://git.io/TVCuCQ | 10:54 |
CIA-11 | shogun: Soeren Sonnenburg master * r7827b83 / (2 files): | 10:55 |
CIA-11 | shogun: add SGVector based dense_dot_vec function wrapping dense_dot to | 10:55 |
CIA-11 | shogun: dotfeatures - http://git.io/EusjMw | 10:55 |
-!- f-x [~user@117.216.153.21] has joined #shogun | 14:16 | |
blackburn | sonne|work: around? | 14:31 |
blackburn | we've got many cache misses in distance and kernel calculation | 14:37 |
blackburn | it is bad to do d[i,j] and d[j,i] in the same time | 14:37 |
blackburn | matrix calculation I mean | 14:38 |
-!- f-x [~user@117.216.153.21] has quit [Remote host closed the connection] | 14:43 | |
-!- f-x [75d89915@gateway/web/freenode/ip.117.216.153.21] has joined #shogun | 14:47 | |
-!- f-x [75d89915@gateway/web/freenode/ip.117.216.153.21] has quit [Ping timeout: 245 seconds] | 14:58 | |
-!- Ram108 [~amma@14.96.24.61] has joined #shogun | 15:23 | |
-!- cronor [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 15:49 | |
sonne|work | blackburn: agreed | 16:24 |
blackburn | I wrote too much to understand what is you agree with :) | 16:25 |
sonne|work | it is bad to do d[i,j] and d[j,i] in the same time | 16:25 |
blackburn | ah yes | 16:25 |
sonne|work | did you write more than that? | 16:25 |
blackburn | well yes, for last two days | 16:25 |
blackburn | ;) | 16:26 |
blackburn | sonne|work: I can hardly wait to release | 16:26 |
blackburn | cause I can't post on mloss and submit paper | 16:27 |
blackburn | blocks me a little | 16:27 |
sonne|work | heikos stuff needs testing before we can | 16:27 |
sonne|work | then do a snapshot | 16:27 |
blackburn | of current version? | 16:27 |
blackburn | how& | 16:27 |
sonne|work | IIRC make release SNAPSHOT=yes | 16:28 |
blackburn | hmm makes sense | 16:28 |
blackburn | sonne|work: have you seen I did integrate covertree? | 16:31 |
blackburn | I know you probably not a big fan of std things but it works smooth, that's definitely better than nothing | 16:31 |
sonne|work | no havent seen it | 16:31 |
sonne|work | no time now | 16:31 |
blackburn | sonne|work: well then you know it now :) | 16:32 |
sonne|work | where is it? | 16:32 |
blackburn | shogun/lib/CoverTree.h | 16:32 |
sonne|work | :) | 16:33 |
-!- Ram108 [~amma@14.96.24.61] has quit [Ping timeout: 248 seconds] | 16:52 | |
@sonney2k | blackburn, what should index_t be? | 16:57 |
blackburn | sonney2k: unsigned? | 16:57 |
@sonney2k | well possible if you are carefull with signedness everywhere | 16:58 |
@sonney2k | would probably be a big change | 16:58 |
blackburn | I think we should -> to unsigned some day | 16:58 |
@sonney2k | blackburn, do you know whether gsoc got announced already? | 16:58 |
blackburn | sonney2k: not yet | 16:58 |
@sonney2k | then things like for int i=0 i< size ; ... won't work | 16:59 |
blackburn | sonney2k: why? | 16:59 |
@sonney2k | we would have to go through all loops and convert them to for index_t i=0 ... | 16:59 |
blackburn | hmm yes, but not a big change | 16:59 |
@sonney2k | some of these are doing plus minus operations like | 17:00 |
blackburn | can be done gradually | 17:00 |
@sonney2k | i-j | 17:00 |
@sonney2k | and check if things are >0 | 17:00 |
@sonney2k | can be done but lot of work | 17:00 |
blackburn | sonney2k: it leaves room for optimization I think, i.e. when compiler knows i is unsigned | 17:00 |
@sonney2k | ? | 17:00 |
@sonney2k | what does it help? | 17:00 |
blackburn | I'm not sure though | 17:01 |
blackburn | don't you think it would be better for compiler | 17:01 |
blackburn | ? | 17:01 |
@sonney2k | underneath (machine language) signed / unsigned is no difference | 17:01 |
blackburn | well, for example loop to zero is faster | 17:02 |
@sonney2k | no | 17:03 |
@sonney2k | same thing | 17:03 |
blackburn | sonney2k: I am not sure, zero test could be faster :) | 17:05 |
blackburn | sonney2k: have you seen my benchmark vs scikits? | 17:05 |
@sonney2k | unsigned int u=0 ====== int u=0 :-) | 17:05 |
blackburn | final one | 17:05 |
@sonney2k | all bits 0 | 17:05 |
@sonney2k | no | 17:05 |
blackburn | you could check if you want | 17:05 |
@sonney2k | where? | 17:05 |
blackburn | in the dropbox paper | 17:05 |
blackburn | folder | 17:06 |
blackburn | damn | 17:06 |
blackburn | :D | 17:06 |
blackburn | in the dropbox folder | 17:06 |
blackburn | now with fruits^W error bars | 17:10 |
CIA-11 | shogun: Soeren Sonnenburg master * ra29c0a6 / (7 files in 4 dirs): fix various doxygen documentation warnings - http://git.io/UOTyDA | 17:18 |
-!- Ram108 [~amma@14.99.70.218] has joined #shogun | 17:20 | |
CIA-11 | shogun: Sergey Lisitsyn master * r2443b79 / (2 files): Added covertree libshogun example - http://git.io/_O90Lg | 17:21 |
blackburn | sonney2k: what is O9 flag? | 17:21 |
blackburn | why not highest O3? | 17:23 |
-!- Ram108 [~amma@14.99.70.218] has quit [Ping timeout: 244 seconds] | 17:44 | |
-!- Ram108 [~amma@14.96.58.242] has joined #shogun | 18:11 | |
-!- f-x [75c0d948@gateway/web/freenode/ip.117.192.217.72] has joined #shogun | 18:27 | |
-!- f-x [75c0d948@gateway/web/freenode/ip.117.192.217.72] has quit [Quit: Page closed] | 18:57 | |
-!- cronor [~cronor@fb.ml.tu-berlin.de] has quit [Ping timeout: 252 seconds] | 19:03 | |
-!- blackburn [~qdrgsm@188.122.239.116] has quit [Quit: Leaving.] | 20:00 | |
-!- blackburn [~qdrgsm@188.122.239.116] has joined #shogun | 20:04 | |
-!- Ram108 [~amma@14.96.58.242] has quit [Ping timeout: 260 seconds] | 20:57 | |
shogun-buildbot | build #134 of nightly_none is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/134 | 21:07 |
CIA-11 | shogun: Soeren Sonnenburg master * r67ea574 / src/shogun/converter/Isomap.cpp : remove unused variables j and tmp from Isomap to fix warning - http://git.io/a8LGnQ | 21:19 |
CIA-11 | shogun: Soeren Sonnenburg master * r9df3f64 / src/shogun/features/BinnedDotFeatures.h : document remaining methods of binned dot features - http://git.io/K8odjg | 21:19 |
CIA-11 | shogun: Sergey Lisitsyn master * r27b603d / examples/undocumented/libshogun/Makefile : Added check-valgrind target for libshogun makefile - http://git.io/-X8Hew | 21:19 |
CIA-11 | shogun: Sergey Lisitsyn master * r03e8d6b / examples/undocumented/libshogun/library_gc_array.cpp : Fixed GCArray libshogun example - http://git.io/yFzyRg | 21:19 |
shogun-buildbot | build #471 of octave_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/471 blamelist: sonne@debian.org | 22:14 |
shogun-buildbot | build #472 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/472 | 22:31 |
CIA-11 | shogun: Sergey Lisitsyn master * r03786e6 / (3 files): Proper automatic k neighborhood matrix handling for LLE,HLLE,LTSA - http://git.io/-bI8ig | 23:58 |
CIA-11 | shogun: Sergey Lisitsyn master * r52334da / src/shogun/converter/KernelLocallyLinearEmbedding.cpp : Covertree max distance computation fix for KLLE - http://git.io/Zfgyhg | 23:58 |
--- Log closed Tue Jan 31 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!