Open in new window / Try shogun cloud
--- Log opened Wed Apr 08 00:00:34 2015
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has joined #shogun00:31
travis-ciit's Soumyajit De'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/5754888900:31
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has left #shogun []00:31
shogun-notifier-shogun: lambday :develop * f4c4bf7 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/f4c4bf74f17fc60f6641e9534246dfe79e4a7f3300:46
shogun-notifier-shogun: Added matrix-type independent add in linalg00:46
shogun-notifier-shogun: Soumyajit De :develop * 701191d / / (3 files): https://github.com/shogun-toolbox/shogun/commit/701191d6ec6bb7703aa1d147cf55aa9e2f4da24200:46
shogun-notifier-shogun: Merge pull request #2811 from lambday/feature/linalg00:46
shogun-notifier-shogun:00:46
shogun-notifier-shogun: Added matrix-type independent add in linalg00:46
shogun-buildbotbuild #282 of trusty - libshogun - viennacl is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/282  blamelist: lambday <heavensdevil6909@gmail.com>00:50
shogun-buildbotbuild #283 of trusty - libshogun - viennacl is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/283  blamelist: Soumyajit De <heavensdevil6909@gmail.com>00:52
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has joined #shogun00:57
travis-ciit's Soumyajit De'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/5756163100:57
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has left #shogun []00:57
-!- PirosB3 [~pirosb3@host33-99-dynamic.181-80-r.retail.telecomitalia.it] has quit [Quit: PirosB3]01:02
@lambdaylisitsyn: there?01:04
lisitsynlambday: hey01:04
@lambdaylisitsyn: how are things? :D01:05
lisitsynlambday: fine01:05
lisitsynwhat about you?01:05
lisitsyncurrently mapreducing some stuff around :)01:05
@lambdaylisitsyn: haha that's great!01:06
@lambdaylisitsyn: I have been working on hypothesis testing a bit01:06
@lambdayand lately (thanks to wu lin) doing some stuffs in linalg as well01:06
lisitsynhmm cool01:06
lisitsynI am on my way to refactor neural nets a bit01:07
@lambdaylisitsyn: one thing I want your opinion on01:07
lisitsynsure01:07
@lambdaylisitsyn: check this benchmark https://github.com/shogun-toolbox/shogun/pull/280601:07
@lambdayit comes from a discussion here https://github.com/shogun-toolbox/shogun/issues/280101:07
lisitsynoh something big01:08
@lambdayso long story short -01:08
@lambdaywu lin was using linalg methods without specifying any backend01:08
@lambdayand he's using SGMatrix01:08
lisitsynokay I've seen there is a bug01:08
lisitsynsome forgotten loading right?01:09
@lambdayso if someone sets viennacl for global backend, then we're doing all the CPU-GPU data transfer01:09
@lambdaywhich stabs the performance at the heart01:09
lisitsynyeah that's obvious01:09
@lambdaylisitsyn: that is fixed - we just needed a variant that returned the result matrix/vector01:10
@lambdayok so what I was thinking is01:10
@lambdayhow about a cmake switch to make SGMatrix as GPU one shogun-wise?01:10
@lambdaywe can just inherit from a separate base01:10
@lambdaybased on some preprocessor flag01:11
@lambdaythat way, we can milk the performance boost to the fullest when GPU is used01:11
@lambdaythis is totally possible even now with explicitly using CGPUMatrix+ViennaCL backed01:11
lisitsynhmm let me think01:12
@lambdaybut how about we use the same SGMatrix everywhere - with the backend changed in a switch01:12
@lambdayI am mainly worried because you see in the benchmark the penalty is unbearable01:13
lisitsynI see no crime in such switch for sure01:13
lisitsynbut I have doubts we can hide gpu01:13
@lambdaybut when we use GPU memory then gain makes me drool :D01:13
lisitsynI mean you always have to care whether you transfer data or not01:13
lisitsynbtw check torch01:14
lisitsynhttps://github.com/torch/torch701:14
@lambdaylisitsyn: checking..01:14
lisitsynneural guys love to use it01:14
@lambdaylisitsyn: what I was saying is something like01:14
@lambdaystruct CPUMatrix01:14
@lambday{01:15
@lambday}01:15
@lambday;01:15
@lambdaystruct GPUMatrix {};01:15
@lambday#ifdef USE_CPU01:15
lisitsynoh you finally talk C++ now I understand ;)01:15
@lambdaylol01:15
@lambdayyou got the idea01:15
@lambdayso two bases01:15
lisitsyncan we get more flexible with less compile time stuff?01:16
@lambdaywhat do you see the problem with that?01:16
lisitsynI am a bit worried about switching everything01:16
lisitsynI mean if sgmatrix is in gpu01:16
lisitsynsome stuff would break right?01:16
@lambdayI'm thinking what would that be..01:17
lisitsynmost of code is not written for gpu so we should expect bugs here01:17
@lambdaywell that makes sense..01:18
@lambdayso devs should explicitly use gpumatrix if they wanna use one..01:19
lisitsynlambday: it seems to me that sgmatrix should be kind of dynamic about that01:19
lisitsynmaybe there is some solution to make it a bit more dynamic01:19
@lambdaylisitsyn: what do you mean by dynamic?01:19
lisitsynlike sometimes it is gpu but sometimes it is cpu01:19
@lambdaylisitsyn: that's what I meant..01:19
@lambday#ifdef USE_CPU01:19
lisitsynI mean not globally01:19
@lambdaylisitsyn: okay.. so some param01:20
lisitsynwell say in neural nets I have all these parameters in sgmatrix01:20
lisitsynthat's cool to have it as gpu matrix01:20
lisitsynbut some other algorithms do not expect gpu01:20
@lambdaylisitsyn: yeah that makes sense...01:21
lisitsynthis doesn't fit your template linalg idea though01:21
@lambdaylisitsyn: what I don't like is that now we have two different types of matrices01:21
lisitsynthat's ok they are really different01:21
lisitsynit seems there is no magic to treat them in the same way01:21
@lambdayyep.. apparently so :(01:22
@lambdayin linalg we're anyway treating them differently01:22
lisitsynyeah01:22
lisitsynI think sgmatrix should be like tensor in torch01:22
lisitsynthey have some clean api it seems01:23
lisitsynbtw what about sgmatrix->sgtensor? ;)01:23
@lambdaylol we do have that multidimensional array thing right?01:23
@lambdaymaybe make that more like tension01:23
@lambdayoops01:23
@lambdaytensor01:23
@lambdaytotally didn't mean tension :D01:24
lisitsynhah01:24
lisitsynthat's actually super cool thing01:24
lisitsynthat they can generalize01:24
lisitsynsay you have RGB image01:24
@lambdayI'm assuming that it is efficient as well01:24
lisitsynand then you got RGB-D and its 3d01:25
lisitsynnothing works01:25
lisitsyn:D01:25
@lambdayvoxels ? :D01:25
lisitsynyeah like these01:25
lisitsynand when aliens get control of humanity01:25
lisitsynwe get to 4d01:25
lisitsynwould be bad to rewrite code again01:26
@lambdaygotta check it then!01:26
lisitsynaha01:26
lisitsynohh these MR tasks go so slow01:27
@lambdayit's in C !01:27
lisitsynyeah lua + C01:27
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has joined #shogun02:18
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has quit [Remote host closed the connection]02:29
-!- lambday [6a33f00b@gateway/web/freenode/ip.106.51.240.11] has quit [Ping timeout: 246 seconds]03:40
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]03:46
shogun-buildbotbuild #995 of nightly_default is complete: Failure [failed notebooks]  Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/995  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, lambday <heavensdevil6909@gmail.com>, Wu Lin <yorker.lin@gmail.com>, Soumyajit De <heavensdevil6909@gmail.com>04:04
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has joined #shogun04:11
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has quit [Remote host closed the connection]04:27
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has joined #shogun05:22
-!- pickle27 [~pickle27@192-0-134-235.cpe.teksavvy.com] has quit [Remote host closed the connection]05:51
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun06:26
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 244 seconds]06:37
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun08:16
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 256 seconds]08:38
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun08:56
-!- HeikoS [~heiko@05453a2b.skybroadband.com] has joined #shogun10:40
-!- mode/#shogun [+o HeikoS] by ChanServ10:40
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 246 seconds]11:34
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun12:11
shogun-notifier-shogun: Wu Lin :develop * f3a5315 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/f3a53156500cd286d3f39e1c3133a86b436b6f9f12:11
shogun-notifier-shogun: added the FITC Laplace method using lbfgs12:11
shogun-notifier-shogun: Wu Lin :develop * 7c1d31f / src/shogun/machine/gp/ (2 files): https://github.com/shogun-toolbox/shogun/commit/7c1d31f9df70e021577c4750d674eedd6a22290212:11
shogun-notifier-shogun: add comments12:11
shogun-notifier-shogun: Wu Lin :develop * 92b55e5 / tests/unit/machine/gp/SingleFITCLaplacianInferenceMethodWithLBFGS_unittest.cc: https://github.com/shogun-toolbox/shogun/commit/92b55e52f12810776b3d6d4f059ebe6d02a8e4a612:11
shogun-notifier-shogun: remove commented codes12:11
shogun-notifier-shogun: Heiko Strathmann :develop * 202951f / / (3 files): https://github.com/shogun-toolbox/shogun/commit/202951f26b6c2e022d7f9cba38a50a06b3b1181812:11
shogun-notifier-shogun: Merge pull request #2804 from yorkerlin/fitc-lbfgs12:11
shogun-notifier-shogun:12:11
shogun-notifier-shogun: Fitc lbfgs12:11
shogun-buildbotbuild #284 of trusty - libshogun - viennacl is complete: Failure [failed test]  Build details are at http://buildbot.shogun-toolbox.org/builders/trusty%20-%20libshogun%20-%20viennacl/builds/284  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Wu Lin <yorker.lin@gmail.com>12:16
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has joined #shogun12:22
travis-ciit's Heiko Strathmann'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/5762202012:22
-!- travis-ci [~travis-ci@ec2-54-226-91-37.compute-1.amazonaws.com] has left #shogun []12:22
shogun-buildbotbuild #3402 of deb1 - libshogun is complete: Failure [failed compile]  Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/3402  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>, Wu Lin <yorker.lin@gmail.com>12:57
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun13:44
-!- HeikoS [~heiko@05453a2b.skybroadband.com] has quit [Quit: Leaving.]14:20
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout]15:11
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 256 seconds]15:56
-!- HeikoS [~heiko@laptop-net.gatsby.ucl.ac.uk] has joined #shogun16:38
-!- mode/#shogun [+o HeikoS] by ChanServ16:38
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun17:55
-!- PirosB3 [~pirosb3@host33-99-dynamic.181-80-r.retail.telecomitalia.it] has joined #shogun18:02
-!- HeikoS [~heiko@laptop-net.gatsby.ucl.ac.uk] has quit [Ping timeout: 244 seconds]20:29
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 244 seconds]20:42
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun22:12
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Read error: Connection reset by peer]22:12
-!- kshitijk [~kshitijk@220.227.149.86] has joined #shogun22:13
-!- kshitijk [~kshitijk@220.227.149.86] has quit [Ping timeout: 246 seconds]22:38
--- Log closed Thu Apr 09 00:00:35 2015