Open in new window / Try shogun cloud
--- Log opened Mon Jul 02 00:00:17 2012
blackburn sonney2k: please do something with R static on buildbot00:06
-!- blackburn [~blackburn@81.28.187.0] has quit [Quit: Leaving.]01:27
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun06:40
n4nd0wiking: around?08:37
-!- uricamic [~uricamic@2001:718:2:1634:7c3b:f22a:ec37:4d] has joined #shogun08:58
-!- cheng [~cheng@115-64-111-17.tpgi.com.au] has joined #shogun09:55
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds]10:02
chengHi all, I spent some time mucking around trying to compile shogun on OSX, and was just wondering whether the problems I'm having are unique to my box.10:07
chengBasically, I have blas/lapack issues.10:07
chengOSX has plenty of different options, and in the end, I had to add --libs=/usr/lib to the configure to make it detect OSX veclib.10:08
chengLots of other options had crazy errors due to gfortran versions, etc10:09
chengJust thought I'd share here to see whether anyone has any experience with this. It seems that /usr/lib should be obvious to add to the default configure.10:10
chengOne more question: how do I get the SVM outputs for binary classification?10:31
chengget_labels returns -1,+1, and I'm not sure how get_confidence works.10:32
-!- cheng [~cheng@115-64-111-17.tpgi.com.au] has quit [Quit: Leaving.]10:32
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun11:11
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun11:33
CIA-18shogun: Sergey Lisitsyn master * r7765f01 / src/shogun/lib/slep/slep_tree_mt_lr.cpp : Fixed tree guided multitask logistic regression - http://git.io/ffIYDA11:39
CIA-18shogun: Sergey Lisitsyn master * r84b7803 / (2 files): Merge branch 'slep' of git://github.com/lisitsyn/shogun - http://git.io/QQh87Q11:39
CIA-18shogun: Sergey Lisitsyn master * r82caba4 / src/shogun/lib/slep/slep_tree_mt_lsr.cpp : Fixes for multitask tree guided least squares - http://git.io/zc_MrQ11:39
blackburnn4nd0: congrats :D11:44
n4nd0blackburn: thank you :)11:44
n4nd0kind of unbelievable, 4-0!11:44
blackburnwell they became pretty upset after 2-011:46
n4nd0Italy?11:47
@sonney2kn4nd0, yeah congrats from my side too. the only game where spain played nicely11:47
@sonney2kthough11:47
@sonney2kwell deserved, very nice goals11:48
n4nd0sonney2k: c'mon ... against Ireland it was good too, similar to yesterday's one to tell the truth11:48
n4nd0the extra time against Portugal was good as well11:48
@sonney2ktrue I forgot about that one11:48
@sonney2kno portugal did actually deserve to win11:48
@sonney2kbut I really expected just a 1:0 from spain11:49
n4nd0:O I don't agree with you there, sorry11:49
@sonney2kand then some spanish wall11:49
n4nd0yeah, they actually played, that was nice11:49
blackburnsonney2k: seen cheng /usr/lib issue?11:50
@sonney2kblackburn, I have no clue what goes wrong with R on the buildbot11:51
@sonney2kI mean it obviously installs ?!11:51
blackburnjust like I have no clue what goes wrong with octave modular :D11:51
@sonney2kblackburn, we need some mac system as buildbot...11:51
@sonney2kblackburn, you mean with sparse typemaps?11:51
blackburnyeah it reproduces only w/o valgrind/gdb11:52
blackburnah no11:52
blackburnit reproduces with gdb11:52
@sonney2kblackburn, btw I can no longer visit github.com since the lates chrome upgrade11:52
@sonney2kalways hangs11:52
blackburnwhen was it?11:52
@sonney2kthe 20.x upgrade11:52
@sonney2kI have no idea what goes wrong11:53
blackburn20.0 here11:53
blackburnworks smoothly11:53
@sonney2kgreat11:53
@sonney2knot here11:53
@sonney2kblackburn, nvidia inside?11:53
blackburnno, some intel11:53
blackburnnot discrete11:53
-!- Netsplit *.net <-> *.split quits: zxtx11:55
-!- Netsplit over, joins: zxtx11:56
@sonney2kno idea - can only tell it sucks :/12:01
blackburnsonney2k: linux world :D12:01
@sonney2kall other browsers work of course12:02
blackburnjust some hardware issue linux is full of12:03
@sonney2kwindoze too but please buy dedicated w* hardware so things are well tested.12:06
@sonney2kn4nd0, btw do things go well now?12:08
blackburnwin is pretty nice with hardware because of $12:08
@sonney2kdid you have a session with uricamic?12:08
n4nd0sonney2k: yes, I am working on the hm-svm12:08
n4nd0I have been discussing with him but nothing special12:08
n4nd0vojtech said he would send me a document with the unconstrained formulation of my problem so I could compare the results of my method12:09
@sonney2kk12:09
n4nd0but I guess he has not managed to find time yet12:09
n4nd0I should talk to Nico soon12:13
n4nd0I hope he has some time today12:13
@sonney2kuricamic, could you please ping vojta about this issue?12:16
n4nd0sonney2k: btw, Nico told me that I'd be able to ask Georg Zeller while he's on holidays, do you know if he's normally available?12:23
@sonney2kn4nd0, just email him12:24
@sonney2kthe sooner the better though12:25
@sonney2kI would even do it this week just to get him to know12:25
@sonney2kI know him personally12:25
@sonney2khe is the other of the matlab hm svm toolbox12:26
@sonney2kand knows all the algorithmics but not so much about shogun12:26
n4nd0ok12:27
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection]12:27
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun12:28
@sonney2kn4nd0, and like nico he is a nice guy so no worries12:28
n4nd0good :)12:28
@sonney2kn4nd0, I am too but that is not so obvious :P12:28
n4nd0hahaha12:29
blackburnyou are not!12:29
blackburn:(12:29
blackburnyou are the dark lord and depression comes around you12:29
@sonney2kblackburn, I know I am bad because I am pressing you to fix the build errors you introduced :P12:37
blackburnmeh :(12:42
blackburnbike time12:42
blackburn:D12:42
-!- blackburn [~blackburn@81.28.187.0] has left #shogun []12:42
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection]12:51
uricamichi, I notified Vojta about that, he promised he will send the document during afternoon12:51
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun12:51
uricamicsonney2k, n4nd0 ^12:51
n4nd0good12:51
n4nd0thank you!12:51
@sonney2kuricamic, thanks!12:51
uricamicno problem, I just returned from the lunch, sorry for delay12:52
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.194.95.174.186] has joined #shogun13:19
-!- cheng [~cheng@115-64-111-17.tpgi.com.au] has joined #shogun13:21
-!- heiko [~heiko@host86-182-167-223.range86-182.btcentralplus.com] has joined #shogun13:23
-!- blackburn [50ea71fe@gateway/web/freenode/ip.80.234.113.254] has joined #shogun13:26
CIA-18shogun: Soeren Sonnenburg master * r8fb2bb1 / src/configure : fix typo in configure HAVE_ARPACK -> HAVE_EIGEN3 - http://git.io/aI9Waw13:27
@sonney2kcheng, we really have problems getting access to some mac hardware13:28
@sonney2kcheng, more permanently that is13:28
@sonney2kwe simply need one to have a buildbot reliably testing that arch13:28
chengsonney2k: I agree.13:29
chengThe general problem with OSX is that one usually has about 4 versions of BLAS/LAPACK.13:30
chengSo a buildbot doesn't really solve that problem.13:30
@sonney2kwell at least we could detect the correct one by default and make sure it stays that way13:31
chengFor example, I have the Xcode version, the enthought python version, the macports version and one more version which I cannot quite figure out why it is there, perhaps gfortran install?13:31
shogun-buildbotbuild #21 of deb1 - libshogun started, including [8fb2bb18892999a00adbd5fafd61da52f075eb0c]13:31
chengAt least on my mac, veclib was detected by adding --libs=/usr/lib13:32
shogun-buildbotbuild #21 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2113:32
shogun-buildbotbuild #29 of deb2 - static_interfaces started, including [8fb2bb18892999a00adbd5fafd61da52f075eb0c]13:32
@sonney2kshogun-buildbot, that was fast13:32
chengsonney2k: how do I get svm scores from BinaryLabels?13:33
@sonney2kcheng, jsut get the confidences13:34
@sonney2kthese are the original outputs of the SVM13:34
chengblackburn: To answer your question via email, adding --libs=/usr/lib makes configure detect veclib correctly and compiles happily.13:34
chengsonney2k: BinaryLabels.get_labels() gives +1, -1.13:34
@sonney2kcheng, get_confidences()13:35
shogun-buildbotbuild #29 of deb2 - static_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/29  blamelist: Soeren Sonnenburg <sonne@debian.org>13:35
shogun-buildbotbuild #28 of deb3 - modular_interfaces started, including [8fb2bb18892999a00adbd5fafd61da52f075eb0c]13:35
* sonney2k *argh*13:35
chengsonney2k: BinaryLabels.get_confidences seems to want an int_32 as an argument.13:35
@sonney2kmy backup machine died13:35
@sonney2kdamn thunderstorms13:35
blackburncheng: I see13:35
@sonney2kblackburn, isn't it get_confidences?13:36
blackburnsonney2k: get_confidences should have no parameter at all13:36
blackburnsonney2k: or what do you ask?13:36
@sonney2kblackburn, yes exactly13:36
@sonney2kcheng, don't mix it up with get_confidence(int)13:37
@sonney2kget_confidences() !13:37
blackburnwhoops I copypasted ARPACK badly13:37
blackburnthanks for fixing13:37
@sonney2kblackburn, heiko did you get an email from the buildbot13:37
@sonney2k?13:37
@sonney2k(just now that is!)13:37
chengblackburn, sonney2k: thanks! Misread the s13:38
blackburnlet me check13:38
chengblackburn: I just tested it. It works.13:39
@sonney2kblackburn, so?13:39
@sonney2kshould be your lisit* gmail address13:39
@sonney2kgtg... nap time is over13:39
blackburnsonney2k: syn13:39
chengsonney2k: later.13:40
blackburnsonney2k: yes I got it13:40
blackburnnew folder at gmail heh13:41
chengblackburn: while I was mucking around with lapack linking, I noted that the GP code need lapack somewhere, but somehow tries to compile it without lapack anyway.13:45
blackburnoh ok I'll let puffin know about it13:45
chengI can try to reproduce that if it helps. Not sure how much it is an OSX think.13:45
chengOSX thing.13:45
blackburnah nevermind it is just a #ifdef13:46
blackburncheng: ah I recall I have jmlr paper :D do you think it makes sense to try again after all that time passed? I've never managed to improve before but could start a little later13:47
chengblackburn: please resubmit soon. :-)13:48
blackburnhow soon?13:48
blackburncheng: you know I wanted to change some code stuff and it would take some time13:49
blackburnanyway I feel pretty unoptimistic about that :D probably will end up with arxiv13:50
chengOf course, the sooner the better, but take your time to do a good job.13:51
chengWhy unoptimistic?13:51
chengpessimistic?13:51
blackburnhah13:51
blackburnyeah13:51
blackburnwell it looks for me just like some crap - I see no valuable part here13:52
chengI see. Well, if it takes longer, the only danger is that you will get new reviewers, who may ask for more stuff.13:53
blackburnah I see13:57
blackburnthen I will try to add more algos here and provide interface13:57
chengIf I remember correctly, the problem was separating from shogun, and not how many algos you have.13:58
blackburnsure13:59
blackburnbut I personally think it lacks some useful algos13:59
blackburnt-SNE in particular, sammon mapping may be14:00
chengYour choice, I am just the messenger between reviewers and authors. :-)14:01
blackburnthank you14:02
shogun-buildbotbuild #28 of deb3 - modular_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/28  blamelist: Soeren Sonnenburg <sonne@debian.org>14:03
heikosonney2k, blackburn,  the buildbot now sends emails to me :)14:08
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Read error: Operation timed out]14:15
-!- cheng [~cheng@115-64-111-17.tpgi.com.au] has quit [Quit: Leaving.]14:16
-!- vojtech [9320543b@gateway/web/freenode/ip.147.32.84.59] has joined #shogun14:54
blackburnheiko: yeah spambuildbot now14:55
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun15:21
-!- nicococo [~nico@lacedcoffee.ml.tu-berlin.de] has joined #shogun15:28
n4nd0hey nicococo15:31
nicococohola n4nd0.. how are you15:32
nicococo?15:32
n4nd0I am good, what about you?15:32
nicococowell, alright, i think you have 4:0 reasons to feel better than me :)15:33
n4nd0haha :D yeah I think so15:34
n4nd0it was a good day yesterday definetely15:34
nicococoi can imagine..15:35
nicococookay, whats going on with the hmsvm?15:35
nicococo(sorry, don't have that much time :( )15:35
n4nd0no problem, I will try to be brief15:35
n4nd0I implemented the joint feature computation15:35
n4nd0and know I am planning how to do the argmax15:35
n4nd0I have almost all the pieces clear for it ... almost :)15:36
nicococofine15:36
n4nd0ok, so what about these vectors p and q?15:37
n4nd0do they need to be estimated using training data?15:37
n4nd0or are they defined in init_opt method15:37
n4nd0?15:37
nicococousually no. in our implementation definitely not.15:38
n4nd0no to which question :P?15:38
nicococoaehm.. they don't belong to the optimization problem of the sosvm, hence they are not defined in the init_op method15:38
n4nd0ok15:39
nicococo2nd question15:39
nicococoare you confused by the p and q.. or is it clear?15:39
n4nd0I don't know how to get them15:40
nicococopre-defined15:40
nicococothink of this: you now in which state you start and in which state you end15:40
n4nd0why?15:41
nicococoand if you don't now.. then add 2 artificial states just for the start and end of each sequence15:41
n4nd0aham15:41
nicococoit has some advantages. and p is mandatory since there is no p(q_1|q_0) defined15:42
nicococodid you read the document?15:42
n4nd0when you say add 2 artificial states15:42
n4nd0do you mean to add two elements to the label sequence15:42
n4nd0but for example, if we have a two-state model15:42
nicococothat would be one possibility15:43
n4nd0these artificial states would be either -1 or 115:43
n4nd0or another new state like 0 or so?15:43
nicococoif you have a 2-state model you need two additional states (liek 3 and 4)15:43
n4nd0ok15:44
nicococobut for now, p and q are just needed in the viterbi15:44
n4nd0yes15:44
n4nd0does that make a difference?15:44
nicococodifference to what?15:45
n4nd0hehe like if to say that they are just needed in the Viterbi makes something different15:45
n4nd0ok, forget about it :)15:45
nicococo:)15:46
n4nd0I think I am not very sure how to handle them, but I can think about it more15:46
nicococoi would suggest to think about it less :)15:47
n4nd0haha ok15:47
nicococoif you look at the definition of the transition matrix15:47
n4nd0yes15:47
nicococoyou have a transition from states y_i-1 to y_i right?15:48
n4nd0yes15:48
nicococowhen you start at the first position of the sequence (i=1) then you have a transition from y_0 to y_1 .. but there is no y_0 hence you need a transition solely depending on the y_115:49
nicococo...and thats p15:49
n4nd0ok, that's all right15:50
n4nd0it's like the pi vector using the conventional notation for hmm I thinks15:50
n4nd0think*15:50
nicococoyes, indeed15:50
n4nd0but in our framework15:51
nicococodid you read the FAQ?15:51
n4nd0yeah :)15:51
nicococois it alright or something missing/unclear?15:51
n4nd0in our framework, where should we define p and q? in the HMMSVMModel?15:51
n4nd0I think that it makes sense there15:51
nicococoi agree15:51
n4nd0yeah, I have a question about the FAQ too15:52
n4nd0it is about the third question15:52
nicococookay15:52
n4nd0so when I said in the question15:52
n4nd0"Then, the15:52
n4nd0first observation sequence, x0 , is never taken into account15:52
n4nd0"15:52
n4nd0that is wrong15:53
n4nd0isn't it?15:53
nicococosince we start with i=1 there is no x_015:53
nicococothe sequences always go from i=1 ... T15:53
n4nd0I said, note that I am talking about 0-index ;)15:53
n4nd0so it's like the sequences go from i=0 ... T-115:54
n4nd0but I think I got the point anyway15:54
n4nd0I have to take into account all the features of x15:54
nicococoaeh. oaky.. sure you can convert them15:54
n4nd0you know ... C people are crazy and start counting the things at zero ... don't blame me :)15:55
nicococoit's okay. i think i can handle that :)15:55
n4nd0I think that the rest is ok15:56
nicococovery good!15:56
n4nd0I am going to do my best to have the Viterbi ready by tomorrow15:56
nicococoso, please try everything out such that if anything is unclear we can talk about it as soon as15:56
nicococopossible15:56
n4nd0sure15:57
n4nd0I can mail you once I have pushed the Viterbi part15:57
nicococothat would be nice.15:57
n4nd0and if you can take a look to the HM-SVM Model it would be great15:57
nicococookay, i can do that. do you think we should talk tomorrow again?15:58
n4nd0maybe we can plan it by mail better?15:58
n4nd0depending on how my progress goes and if I get lot of doubts and stuff15:58
nicococoalright. so, if question, hints, or whatever please mail me and we can add everything to the document15:59
n4nd0cool15:59
n4nd0ok, thank you very much then!15:59
n4nd0good luck with your work ;)15:59
nicococoso, very good work, thank you for that.. and see you soon15:59
nicococothanks :)15:59
-!- nicococo [~nico@lacedcoffee.ml.tu-berlin.de] has left #shogun []16:00
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has left #shogun []16:01
-!- blackburn [50ea71fe@gateway/web/freenode/ip.80.234.113.254] has quit [Quit: Page closed]16:18
-!- vojtech [9320543b@gateway/web/freenode/ip.147.32.84.59] has quit [Quit: Page closed]17:14
-!- heiko1 [~heiko@host86-177-172-27.range86-177.btcentralplus.com] has joined #shogun17:16
-!- heiko [~heiko@host86-182-167-223.range86-182.btcentralplus.com] has quit [Ping timeout: 264 seconds]17:18
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun17:27
puffin444Hey heiko17:30
heiko1hey puffin44417:38
puffin444what do you think about the model selection checksum?17:38
heiko1I actually just started reading17:39
heiko1Ill get back to you when I understoof it :)17:39
puffin444ok17:40
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun17:43
n4nd0sonney2k: hey there17:44
heiko1puffin444 this get_char_description17:49
heiko1what do you need it for? the hashing?17:49
puffin444Yes. I use it to make a char array that the MD5 function can hash. Do you think its too complicated?17:50
heiko1it looks a bit strange, why all these type distinctions and redundant code?17:50
puffin444I store the value of the parameter in the description; the value may need more space depending on the type.17:51
heiko1oh btw there is a method to get the size of a TParameter's primitive type17:51
puffin444That would help alot.17:51
heiko1its a bit hidden: have a look at TSGDataType in DataType.h17:53
heiko1every TParameter instance has a TSGDataType17:53
heiko1and this you can ask these things17:53
heiko1but I still dont really get these loops17:53
puffin444Yes that would simplify things.17:54
heiko1big_name contains a char description of the data of the parameter?17:54
puffin444big_name contains a description of the name and values of the parameter as well as all of its children.17:55
puffin444The char description is recursively built. In order to store a value, I cast it as an unsigned char and store it that way. That is what the loops are for.17:56
heiko1where children are respect to what?17:56
puffin444Children that are SGObjects might have modelselection parameters themselves. I make sure that they are part of the description.17:56
heiko1ok, i see ...17:57
heiko1continuing reading ....17:57
heiko1why only model-selection parameters?17:57
heiko1because you only want to know when these changed, right?17:58
puffin444Yes. I suppose I could check all parameters with m_parameters, yes?17:58
heiko1but are other parameters not als important?17:58
heiko1for what do you need that hashing?17:58
heiko1for model-selection?17:58
heiko1or for just knowing when to recompute stuff?17:59
puffin444Knowing when to recompute stuff. sonney2k suggested it so I tried it out.17:59
heiko1ok, yeah I mean its a nice idea17:59
heiko1but I would include all parameters18:00
puffin444will do.18:00
heiko1since these might change during training or whatsoever18:00
heiko1still this description code is strange ;)18:00
puffin444Yes I did fear it was a little over-complicated.18:00
heiko1I will go thorugh it a bit more and see if its possible to simplify18:03
puffin444I also don't know if it's really appropriate for the parameter itself to include SGObject children in its description, but I don't think this checksum would be useful if I didn't look at all of parameters.18:03
heiko1no youre right, a parameter has changed if the sub-parameter has changed18:04
heiko1imagine a svm with a kernel with width18:04
heiko1when the width has changed, the kernel has changed and an evtl kernel matrix in svm has to be recomputed18:04
heiko1puffin444 I think this has to be done differently, see comments in pull request18:24
puffin444Absolutely18:25
puffin444I have been reading your comments.18:25
heiko1(but I dont know how currently, complex thing)18:25
heiko1I also just added some thoughts on the general approach18:25
puffin444heiko, perhaps murmur would be better?18:31
puffin444If I interpret this correctly, I can call CHash::IncrementalMurmurHash incrementally18:32
heiko1yeah could do that18:33
heiko1you would need to extend it a bit18:33
heiko1so that you can specify an address and a number of bytes18:33
puffin444sounds like a simple thing to do.18:34
heiko1yeah, the more complicated thing is not to loose the overview :)18:34
heiko1have a look at the serialisation code first18:34
heiko1thats pretty complicated18:34
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun18:35
heiko1many cases18:35
puffin444Doesn't serialization lead to the same problem? I'm doing the copying again, instead this time its for something to be written on disc.18:36
heiko1yes18:37
heiko1you can write to memory18:37
heiko1ybut i meant the way the parameters are visited18:37
heiko1because its exectly the same problem: print/hash values of all parameters (including sub-parameters, vectors/matrices etc)18:38
puffin444Yes. I can't find the class that does this. What's its name?18:39
heiko1SGObject starts the process18:39
heiko1save_serializable or so18:39
puffin444Oh now I see.18:39
heiko1I meant just for inspiration18:40
@sonney2kn4nd0, yes?18:40
heiko1you dont need to go into everything, just for getting a feeling for all the cases18:40
@sonney2kblackburn, well I always received these emails -guess why I spend so much time fixing issues others create?18:40
CIA-18shogun: Heiko Strathmann master * re5b0dbe / (2 files): removed some of the asserts which might be too strong - http://git.io/nMRneQ18:40
CIA-18shogun: Heiko Strathmann master * rfea2a1e / examples/undocumented/libshogun/Makefile : added statistics examples again (hopefully work now) - http://git.io/FVOQ1Q18:40
CIA-18shogun: Heiko Strathmann master * rf72cc6d / (3 files): Merge pull request #616 from karlnapf/master - http://git.io/b4I9lQ18:40
blackburnwhat?18:40
puffin444Yes this looks like a much better approach.18:41
shogun-buildbotbuild #22 of deb1 - libshogun started, including [e5b0dbeead7748f512a081d326cbb7e388c2b774]18:41
heiko1puffin444, if you then iterate over all these parameters and incrementally hash18:41
heiko1then everything is in place and nice18:41
heiko1also I would moive the stuff to TParameter as I wrote18:42
shogun-buildbotbuild #22 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2218:42
shogun-buildbotbuild #23 of deb1 - libshogun started, including [fea2a1e8af7bc6d513d3550705ba5bfdc184ae11, f72cc6d6bb058427d603f016ab9898c2ecf66f07]18:42
heiko1to be able to hash single parameters if needed from another place18:42
puffin444will do.18:42
blackburnsonney2k: what is this lament about? :D18:42
heiko1and then a method hash_all_parameters or so for SGObject18:42
shogun-buildbotbuild #23 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2318:42
shogun-buildbotbuild #30 of deb2 - static_interfaces started, including [e5b0dbeead7748f512a081d326cbb7e388c2b774]18:42
puffin444I am also working on integrating vector parameters into model selection18:42
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Remote host closed the connection]18:43
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun18:43
heiko1wow, another large thing :)18:43
heiko1you will be a framework guru when gsoc is over :D18:43
shogun-buildbotbuild #31 of deb2 - static_interfaces is complete: Retry [retry exception slave lost]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/31  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:43
shogun-buildbotbuild #32 of deb2 - static_interfaces started, including [e5b0dbeead7748f512a081d326cbb7e388c2b774, fea2a1e8af7bc6d513d3550705ba5bfdc184ae11, f72cc6d6bb058427d603f016ab9898c2ecf66f07]18:43
n4nd0sonney2k: can you open this http://83.49.221.169:8080/home?18:44
heiko1argh18:44
n4nd0it is a very first draft though :D18:44
blackburnhah18:44
blackburnI have seen that template somewhere18:44
@sonney2kn4nd0, heh - only problem is that this is the style I use for my homepage - so I would rather have used that other one...18:44
heiko1sonney2k, what is that?18:45
heiko1remoteFailed: [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionLost'>: Connection to the other side was lost in a non-clean fashion. ]18:45
@sonney2k...I pointed too :/18:45
@sonney2kheiko1, I am still messing with the bb18:45
heiko1sonney2k, ah ok puh :)18:45
@sonney2kheiko1, trying to figure out why email notifications for dependent builds don't work18:45
heiko1oh man18:46
heiko1sounds annoying18:46
@sonney2kheiko1, my only luck is that people on #buildbot are *extremely* helpful18:46
@sonney2klike here I would say18:46
heiko1ah thats good18:46
heiko1power of open-source :)18:46
shogun-buildbotbuild #32 of deb2 - static_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/32  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>18:47
shogun-buildbotbuild #29 of deb3 - modular_interfaces started, including [e5b0dbeead7748f512a081d326cbb7e388c2b774, fea2a1e8af7bc6d513d3550705ba5bfdc184ae11, f72cc6d6bb058427d603f016ab9898c2ecf66f07]18:47
-!- uricamic [~uricamic@2001:718:2:1634:7c3b:f22a:ec37:4d] has quit [Quit: Leaving.]18:47
puffin444heh lol18:48
@sonney2kheiko1, blackburn btw I increased verbosity of the bb to maximum18:48
@sonney2kbasically telling us when it builds what18:48
@sonney2kwhen it finishes with error/success18:48
@sonney2ketc18:48
heiko1sonney2k, its nice to see when it start/ends18:48
blackburnsonney2k: the only missing thing is bettyboo-style chatting18:48
@sonney2khopefully18:48
@sonney2kheiko1, I unfortunatelly killed the other buildslaves at TU and the admin is on vacations18:49
blackburnkilled admin and buildslaves are on vacations?18:49
@sonney2kactually 'killed' windows did some security upgrades and did restart18:49
heiko1lol :)18:50
heiko1ah the joys of windows18:50
@sonney2kso now it is gone18:50
heiko1man, hopefully people wont kill you :D18:50
@sonney2kheiko1, blackburn opinions on http://83.49.221.169:8080/home?18:50
@sonney2kwhat n4nd0 posted...18:51
blackburnsonney2k: I thought it is your homepage first18:51
blackburn:D18:51
@sonney2kI personally would choose a different style for obvious reasons18:51
@sonney2kit is maybe not such a bright idea to connect shogun too much with myself18:51
n4nd0yeah ... sure18:51
@sonney2kI dont' want this project to end if I loose interest or so18:52
n4nd0the guy who is working on it will appear around soon and ask which one you like18:52
heiko1sonney2k, I like the tabs18:52
heiko1design - dont know18:52
heiko1I dont like the archive links18:52
blackburnsonney2k: hahah what do you expect if you loose interest then18:53
heiko1they are messy and people are probably not interested in that18:53
heiko1actually I would remove this sidebar18:53
heiko1I like the categories18:54
heiko1What would be nice would be a feature list, which is a bit nicer to overview that the main text18:54
heiko1gotta catch train, see you later18:55
@sonney2kblackburn, there are enough people nowadays that know their way around in shogun18:57
blackburnwe need stalin here :D18:58
blackburnsonney2k: any stalindates?18:58
@sonney2kblackburn, stalin is the national hero of some other $COUNTRY18:58
blackburnnational what?18:58
blackburn:D18:58
blackburnsonney2k: he can be hero only of some really braindamaged people18:59
-!- heiko1 [~heiko@host86-177-172-27.range86-177.btcentralplus.com] has quit [Ping timeout: 264 seconds]19:00
blackburnsonney2k: ok in other words we have no leader but you - better?19:00
@sonney2kn4nd0, did you see heikos feedback?19:01
@sonney2kn4nd0, ok cool19:01
n4nd0sonney2k: I'll tell my friend to read it in the logs19:02
-!- bern4rd [~bernard@169.Red-83-49-221.dynamicIP.rima-tde.net] has joined #shogun19:12
@sonney2kn4nd0, I had this template in mind http://www.nuviotemplates.com/template/gr8-01/19:12
@sonney2khi bern4rd19:12
bern4rdhi19:12
@sonney2kinferring from the 4 in you name you are a friend of n4nd0 :D19:12
bern4rdyes :)19:13
blackburnhaha19:13
blackburnreally?19:13
@sonney2kbern4rd, so we had some brief discussion - basically we have a webtemplate clash (the one you used is also the one I use on my homepage and shogun should not be too connect to me)19:14
@sonney2kso I would prefer http://www.nuviotemplates.com/template/gr8-01/19:14
bern4rdi am downloading it right now19:14
@sonney2kbern4rd, is the page already django based or did you 'just' modify the template btw?19:15
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 264 seconds]19:15
@sonney2kbern4rd, I just now realize that one can already click through things19:15
bern4rdthe template is based in django19:15
@sonney2kso you did all the website work already !19:16
shogun-buildbotbuild #29 of deb3 - modular_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/29  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>19:16
bern4rdthe basic stuff yes19:16
bern4rdbut i suppose you want to add more things19:17
bern4rdmaybe a blog to allow your developers comment their work19:17
bern4rdright now the webpage is only showing content so is not too difficult19:18
@sonney2kbern4rd, but that is all not so important compared to the redesign19:18
@sonney2kbern4rd, so the admin users can edit this content right?19:18
bern4rdyes yes I know, i will use the template gr8-0119:18
@sonney2kbern4rd, no what I meant is going from the old page to something new19:19
@sonney2knot the cry* style to gr8 change19:19
@sonney2kbern4rd, I am impressed already how much work you invested...19:19
bern4rdthanks :)19:20
bern4rdas we set the deadline on wednesday can I continue working on it?19:21
@sonney2kbern4rd, sure sure19:21
@sonney2kbern4rd, just come here for updates and discussions19:21
bern4rdsure19:22
@sonney2kI would say you convert to gr8-01 first19:22
@sonney2kthen we finalize tabs19:22
@sonney2kand then we can actually already put it live19:23
bern4rdnice19:23
@sonney2kblackburn, did you for build failure #29 receive an email?19:24
blackburnsonney2k: I received only one19:25
blackburnno idea about #19:25
@sonney2kblackburn, this one <shogun-buildbot> -build #29 of deb3 - modular_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/29  blamelist: Heiko Strathmann <heiko.strathmann@gmail.com>19:26
@sonney2k?19:26
@sonney2kbern4rd, thanks for you work19:26
blackburnBuild Reason: downstream19:26
blackburnBuild Source Stamp: [branch master] 8fb2bb18892999a00adbd5fafd61da52f075eb0c19:26
blackburnBlamelist: Soeren Sonnenburg <sonne@debian.org>19:26
blackburnsonney2k: just check 'To'19:26
bern4rdyou're welcome19:26
bern4rdI go to work. see you19:27
-!- bern4rd [~bernard@169.Red-83-49-221.dynamicIP.rima-tde.net] has quit [Quit: Lost terminal]19:27
blackburnsonney2k: argh19:28
blackburnspam catched it19:28
blackburnI did19:28
blackburnsonney2k: no idea what is wrong with octave modular19:36
blackburnsonney2k: PARALLEL19:44
@sonney2kwhat is run in parallel?19:46
@sonney2kblackburn, with directors or w/o?19:47
blackburnsonney2k: get_feature_vector19:47
blackburnoperator= running in parallel does BOOM19:47
blackburnsonney2k: I have a few possible solutions in mind19:47
blackburndisable parallel for sparse, synchronize at sparsefeatures or synchronize ref in sgrefdata19:48
blackburnsonney2k: your choice?19:48
@sonney2kblackburn, it should crash not only in octave19:50
@sonney2kwe indeed have a couple of algos that are running in parallel for sparse stuff19:50
blackburnsonney2k: does distance_sparseeuclidean_modular.py work on your machine?19:51
@sonney2kno idea19:51
blackburnI am pretty sure it fails in octave for this reason19:52
blackburnonce I disable threads it works19:52
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.194.95.174.186] has quit [Ping timeout: 245 seconds]19:54
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Ping timeout: 245 seconds]19:55
@sonney2kblackburn, well this example should compute the distance matrix in parallel19:56
@sonney2kthough I see conflicts only on diagonal19:56
blackburnsonney2k: so?19:57
blackburnwe need synchronization here19:57
@sonney2kyes19:57
blackburnwhere?19:57
@sonney2ksame overhead as in SGObject19:57
blackburnI'll add it to sparsefeatures then19:58
@sonney2kphtread_mutex_t is 40 bytes20:01
@sonney2kspinlock is just int20:01
@sonney2kso 4 bytes20:01
@sonney2kso 4 more bytes are OK I would say20:02
blackburnsonney2k: I'll add it to get_feature_vector then20:02
@sonney2kblackburn, so do the same as in sgobject in sgreferenced data20:02
blackburnnot to sgrefdata20:02
blackburndo you think so?20:02
blackburnI don't mind any20:03
@sonney2kyeah, we will run into other issues and 4 bytes per vector is something we have to live with20:03
@sonney2kyes shogun won't be fastest this way when we have 3 dimensional data sets20:03
@sonney2kbut why would one need billions of examples to get accurate estimates for such low dim data anyways?20:04
@sonney2kso we are now at 16 byte overhead20:04
@sonney2kand if we are really unhappy with it we can of course change things later20:05
blackburnsonney2k: argh what is behavior on =?20:07
blackburnthere should be pointer on lock I guess?20:07
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun20:09
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Client Quit]20:10
@sonney2kblackburn, ???20:12
@sonney2kblackburn, whenever the refcount stuff is changed we need the spinlock20:12
blackburnsonney2k: just like reference counter it should be shared between vectors/matrices20:12
@sonney2kahh20:12
@sonney2kkaboom20:12
blackburnwhy so?20:13
@sonney2kok maybe it is not too bad20:13
@sonney2kjust create a struct with the refcount and spinlock20:14
@sonney2kand alloc this in SGReferencedData20:14
blackburnwhy struct?20:14
@sonney2kso the former refcount ptr points there20:14
@sonney2kblackburn, how could it work otherwise?20:14
@sonney2kyou need a lock for the refcount20:14
@sonney2kand you need to share it among SGVectors20:15
blackburnsonney2k: just track both separately20:15
@sonney2kwhy that?20:15
@sonney2kthen you have 2 ptrs and 2 ints20:15
@sonney2kwith a struct you will only have 1 ptr and 2 ints20:15
blackburnahh20:15
@sonney2k8 bytes saved20:15
blackburnok makes sense20:15
@sonney2kI hope there is no race condition accessing *ptr though20:16
blackburnwhoops20:16
blackburnhah20:16
blackburnwell20:16
blackburnlets see20:16
blackburnI don't think there is20:16
blackburnsonney2k: I actually have pointers to locks in LTSA/LLE/blahblah20:17
blackburnworks nice I believe20:17
@sonney2kyeah lets hope so20:17
@sonney2kactually it should be fine... only race condition would be when the struct is destroyed20:18
@sonney2kbut this can only happen when only one object is available20:18
@sonney2kso all good20:18
@sonney2klet me break the build just for the fun of it20:19
blackburnsonney2k: ?20:19
blackburnI am doing that now20:19
@sonney2kblackburn, you continue fixing things20:20
@sonney2kI am testing the bb20:20
blackburnah20:20
blackburnI thought you want to break build with these changes20:20
CIA-18shogun: Soeren Sonnenburg master * rebfa8bf / src/configure : breaking configure (buildbot test) - http://git.io/Hxu-4g20:21
@sonney2kblackburn, I am configuring the bb for the last several days so no20:21
@sonney2konce I am done I focus on website20:21
shogun-buildbotbuild #24 of deb1 - libshogun started, including [ebfa8bf667745650d24afd845ccb239d8d43ffd6]20:23
shogun-buildbotbuild #24 of deb1 - libshogun is complete: Failure [failed configure]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/24  blamelist: Soeren Sonnenburg <sonne@debian.org>20:23
@sonney2kblackburn, hah perfect I got an email to my gmail and debian address :-))20:25
@sonney2know lets make things a bit more complicated...20:25
blackburnI didn't20:25
blackburndamn gmail20:25
blackburnhmm20:26
CIA-18shogun: Soeren Sonnenburg master * rcf9d5de / src/configure : Revert "breaking configure (buildbot test)" - http://git.io/fHBEOw20:26
blackburnsonney2k: should I receive it?20:26
@sonney2kblackburn, yes20:26
@sonney2kand I've seen you got it20:26
@sonney2k(the email had you in CC)20:27
blackburnidiot gmail keeps putting it into spam20:27
blackburnnow in main account20:27
shogun-buildbotbuild #25 of deb1 - libshogun started, including [cf9d5deaa537aad24b403053e88aede97d6fad13]20:28
@sonney2khmmhh no idea how to fix that :/20:28
blackburnsonney2k: fix whaT?20:28
@sonney2kblackburn, that gmail puts it into spam20:28
shogun-buildbotbuild #25 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2520:29
shogun-buildbotbuild #33 of deb2 - static_interfaces started, including [cf9d5deaa537aad24b403053e88aede97d6fad13]20:29
blackburnah well I set up filter20:29
@sonney2khaha success email arrived :D20:29
blackburnforcing it not to go to spam20:29
@sonney2know lets see how it does about static20:29
blackburnyeah20:29
blackburnhere20:29
@sonney2kblackburn, btw ccache is on the bb20:29
blackburnohh that's nice20:29
@sonney2kso that is why it took only 1 min20:30
@sonney2kor less :D20:30
blackburnsonney2k: I really would like to see interfaces compiling faster20:30
@sonney2kme too20:30
@sonney2kbut I don't know how20:30
@sonney2kblackburn, btw where is gsomix?20:31
blackburnsonney2k: I don't know20:31
blackburnsonney2k: ++(m_refcount->rc) - does this work?20:31
@sonney2ksure20:31
shogun-buildbotbuild #33 of deb2 - static_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/33  blamelist: Soeren Sonnenburg <sonne@debian.org>20:32
shogun-buildbotbuild #30 of deb3 - modular_interfaces started, including [cf9d5deaa537aad24b403053e88aede97d6fad13]20:32
-!- blackburn [~blackburn@81.28.187.0] has left #shogun []20:32
@sonney2kyes it works :D20:33
-!- blackburn [~blackburn@81.28.187.0] has joined #shogun20:36
@sonney2klet me disassemble the backup machine20:37
* sonney2k prays everything will be OK once this is done20:37
blackburn+86C20:37
blackburnuhh20:37
@sonney2kheh20:37
blackburnultrabooks are really hot20:37
blackburnmodshogun_wrap.cxx: In function 'octave_value_list _wrap_refcount_t_lock_get(const octave_value_list&, int)':20:39
blackburnmodshogun_wrap.cxx:76885:124: error: invalid conversion from 'volatile void*' to 'void*' [-fpermissive]20:39
blackburnsonney2k: ooop!20:39
blackburnvolatile void??20:39
@sonney2kwell we don't need that function right?20:40
@sonney2kso %ignore it20:40
@sonney2kand the data type20:40
@sonney2kbut yes volatile seems necessary20:40
blackburnsonney2k: I do not understand why it appears20:41
@sonney2kblackburn, swig ...20:41
-!- gsomix [~gsomix@188.168.2.163] has joined #shogun20:41
blackburnahh so it added new struct?20:41
gsomixgood evening20:41
gsomixsonney2k, around?20:41
gsomixsorry I was gone .__.20:42
@sonney2kgsomix, hey there20:42
@sonney2kblackburn, please just use %ignore for the struct and lock functions20:42
gsomixsonney2k, I have some problems with directors.20:42
blackburnsonney2k: isn't IGNORE_IN_CLASSLIST enough?20:42
@sonney2kno need to have that in interfaces20:42
@sonney2kblackburn, that is only for serialization20:42
blackburnargh20:42
blackburnok20:43
@sonney2kgsomix, I also have problems with directors20:43
gsomixhuh20:43
gsomixmy problems related to dotfeatures20:43
gsomix*is20:44
@sonney2kgsomix, don't stop there20:44
gsomixmmm? but this is important20:44
@sonney2kgsomix, == please continue20:45
gsomixah, ok20:45
@sonney2kgsomix, I am waiting for a proper problem description20:45
@sonney2kso far there is no way anyone can help20:45
blackburnargh still fails with refs locking :(20:46
@sonney2kblackburn, hmmh - but it works with nthreads = 120:49
@sonney2k??20:49
blackburnsonney2k: wait was typo here20:49
@sonney2kheh20:49
blackburnhmm20:49
blackburnno20:49
blackburnstill20:49
@sonney2kblackburn, with threads == 1 all good?20:49
blackburnsonney2k: yeah..20:50
@sonney2kblackburn, can you do one test20:51
@sonney2kincrease refcount of these SGReferenced data stuff to say 1000 or so20:51
@sonney2kif the problem still appears then it must be sth else20:52
blackburncompiling20:53
gsomixOk, we have add_to_dense_sgvec(blabla, SGVector<float64_t>, blabla) method. If this method is called form C++ side, we get SGVector object on python side. Instead numpy array.20:54
gsomixSGVector have `+=' and `+' operators, so:20:54
gsomixSGVector+=numpy_array - works proper20:54
gsomixSGVector+=SGVector - fail, because SGVector is not numpy array, as required by typemaps.20:54
gsomixI tried to use directorin/out typemaps, but it's fail. It seems there is bug in swig2.0.4. Next I installed swig2.0.7 and got segfault in compilation.20:54
gsomixAnd last. I tried to use directored features in svm - fail:20:54
gsomix>> dsvm.train()20:54
gsomixTypeError: not a numpy vector of appropriate type20:54
gsomixI think it's related to method with SGVector<>.20:54
gsomixsonney2k, that's all. so floody .___.20:54
gsomix*related to methods20:55
blackburngsomix: what is svm you have been using?20:56
gsomixblackburn,  LibLinear20:58
blackburnok20:58
@sonney2kgsomix, python buffer protocol certainly would workaround that21:00
shogun-buildbotbuild #30 of deb3 - modular_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/30  blamelist: Soeren Sonnenburg <sonne@debian.org>21:01
shogun-buildbotbuild #6 of nightly_none started, including []21:01
@sonney2kgsomix, btw as workaround one should be able to create a SGVector from numpy and so then add things21:02
gsomix>> create a SGVector from numpy | on python side?21:02
gsomixbut typemaps do not allow such a method.21:03
@sonney2kgsomix, v=SGVector(numpymatrix)21:05
@sonney2kI think it is called RealVector(numpymatrix)21:05
@sonney2kblackburn, yay! buildbot email notification works now :))21:06
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.194.95.174.186] has joined #shogun21:06
blackburnsonney2k: bah.. that didn't help21:10
blackburnokay so I have no idea then21:11
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun21:13
shogun-buildbotbuild #6 of nightly_none is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/621:13
shogun-buildbotbuild #9 of nightly_default started, including []21:13
blackburnsonney2k: when I added debug print it worked21:18
blackburn:D21:18
blackburnheisenbug21:18
@sonney2kblackburn, debug prints are kind of sleep delays...21:27
blackburnyeah I think I know what is wrong21:28
@sonney2k?21:28
blackburnI didn't synchronized deletion21:28
blackburnso one sets to 021:28
blackburnand other sets to -121:28
blackburnbut both try to delete21:28
blackburnmy guess that is21:29
-!- alexlovesdata_ [c25faeba@gateway/web/freenode/ip.194.95.174.186] has quit [Quit: Page closed]21:29
blackburnyeah21:29
blackburnworks21:29
gsomixsonney2k, ok, thanks. I'll try it tomorrow.21:30
gsomixsonney2k, so what about swig2.0.7? segfault looks strange21:31
blackburnwe should help swig and report about segfault I think :)21:32
blackburnhowever that is not swig21:32
blackburnit is g++ breaks, right?21:32
gsomixyep21:32
blackburnhuh21:32
blackburng++ dev team would be happy to try to recognize the error in 300K LoC file21:33
blackburn:D21:33
blackburnhowever I know one way I think21:33
blackburnone need to compare modshogun.cxx produced by swig 2.0.421:34
blackburnand by swig 2.0.721:34
gsomixaha21:35
gsomixwow, thunderstorm21:41
gsomixit's good that I'm in the big city now :)21:41
shogun-buildbotbuild #9 of nightly_default is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/921:46
shogun-buildbotbuild #5 of nightly_all started, including []21:46
blackburnshogun-buildbot: work21:48
CIA-18shogun: Sergey Lisitsyn master * rf1a01b7 / (3 files in 3 dirs): Made refcounting in SGReferencedData thread-safe - http://git.io/SieMkA21:48
-!- hoijui [~hoijui@dslb-088-074-106-085.pools.arcor-ip.net] has joined #shogun22:07
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun22:10
n4nd0gsomix: what kind of news we will have to today? :)22:13
gsomixn4nd0, do you mean that I should to write weekly report? okay. :]22:15
gsomixall news in log22:15
n4nd0haha no no22:15
n4nd0I don't have kick-ass mode22:15
n4nd0:P22:15
gsomixbut I should :)22:16
n4nd0I am a fan of your comments type "good news everyone"22:16
n4nd0for real, I think they are original :)22:16
@sonney2kblackburn, dammed looks like my backup machines power supply died22:17
@sonney2klikely the thunderstorm caused this22:17
blackburnwhat is wrong with it?22:17
@sonney2kblackburn, turn compu on 'tsss' noise -> off22:17
shogun-buildbotbuild #5 of nightly_all is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/522:18
shogun-buildbotbuild #26 of deb1 - libshogun started, including [f1a01b70ec1519a0d586f5588d55b9927df84fc7]22:18
blackburnsonney2k: what?!22:18
@sonney2kblackburn, ?22:18
blackburnsonney2k: so it produces no voltage or what?22:18
@sonney2kwell for <1 second it does22:19
@sonney2kcould be that the PSU is OK and the board or whatever is RIP22:19
@sonney2kbut how can I find out :/22:19
@sonney2kI have to buy a new one to find the cause22:19
@sonney2kand most likely is the psu22:20
shogun-buildbotbuild #26 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2622:20
shogun-buildbotbuild #34 of deb2 - static_interfaces started, including [f1a01b70ec1519a0d586f5588d55b9927df84fc7]22:20
blackburnsonney2k: measure voltage of PSU runned separately22:20
gsomixn4nd0, I'll try to `repass' an exam tomorrow. This will increase my uni stipend for 10$ per month. :D22:21
n4nd0gsomix: cool, I wish you good luck with it. What is the subject?22:21
gsomixoptics, optics, again optics22:21
gsomixtnx :)22:22
n4nd0you must be an expert on the topic22:22
blackburnsonney2k: just shortcircuit black and green or so (shoold be googleable) :)22:22
gsomixn4nd0, this is my future work, hehe22:22
@sonney2kblackburn, :P22:23
blackburnsonney2k: seriously - I did that twice or so22:23
@sonney2k????22:24
blackburnjust run your PSU w/o anything and you will get whether it works22:24
@sonney2kblackburn, how so?22:24
@sonney2kit dies under load22:24
blackburnsonney2k: are you sure it dies under load?22:24
@sonney2kblackburn, yes22:24
blackburnhmm22:25
@sonney2kat least the mainboard led is on / network led too afterwards22:25
@sonney2kso the board seems to be OK22:25
@sonney2kbut who knows...22:25
@sonney2kblackburn, I was just looking at your patch22:25
blackburnand?22:25
@sonney2kI am a bit concerned that init_data() is called in a pthread lock thing22:26
blackburnwhy?22:26
@sonney2k(which is a busy loop)22:26
@sonney2kbetter store the refcount before22:27
blackburninit_data is basically two 'mov' in all our structures22:27
@sonney2kthen free etc22:27
blackburnfree should be inside LOCK22:27
@sonney2kno22:27
blackburnwhoops22:28
blackburnwrong here for sure22:28
@sonney2kI mean if it is 022:28
@sonney2kthen there really is no one else fiddling with the object22:28
@sonney2knot now and cannot be in future22:28
blackburnsonney2k: bug here anyway22:28
blackburnreturn before unlock22:28
@sonney2kyes22:29
@sonney2kbetter do int c = m_refcount->rc; directly after the decrement22:29
@sonney2kand then unlock22:29
@sonney2kand do whatever necessary22:29
blackburnI feel too lazy to think about it22:29
blackburnwill just check whether it works22:29
@sonney2kblackburn, err what is line 107ff doing?22:30
blackburnsonney2k: so you mean in thready case rc will never be zero?22:30
@sonney2kshouldn't it be ifndef ?22:30
blackburnoops22:30
@sonney2kblackburn, no why?22:31
blackburnsonney2k: I am concerned with free out of pthread lock22:31
@sonney2kblackburn, if the refcount is 022:32
@sonney2kthen we free right?22:32
blackburnyes22:32
@sonney2kso it means that this is the only object still available22:33
blackburnsonney2k: but how do we guarantee we free only once?22:33
@sonney2kno one else can have one22:33
blackburnlooks like22:33
blackburnsonney2k: what if we have two copies22:35
blackburnref = 222:35
blackburnthey both decrement ref22:35
blackburnto zero22:35
@sonney2kthat cannot happen22:35
@sonney2kone will decrement first22:35
@sonney2kso only one of them can call free22:35
blackburnsonney2k: should be better now22:40
CIA-18shogun: Sergey Lisitsyn master * r8bc0fdf / src/shogun/lib/SGReferencedData.h : Updated references thread-safe handling - http://git.io/22qlmw22:40
shogun-buildbotbuild #34 of deb2 - static_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/34  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>22:42
shogun-buildbotbuild #31 of deb3 - modular_interfaces started, including [f1a01b70ec1519a0d586f5588d55b9927df84fc7]22:42
blackburnare you kidding??!22:44
@sonney2k?22:46
blackburnsonney2k: it hangs on classifier_libsvm22:46
-!- hoijui [~hoijui@dslb-088-074-106-085.pools.arcor-ip.net] has quit [Quit: Leaving]22:54
@sonney2kgsomix, when is your exam tomorrow?23:09
-!- naywhaya1e is now known as naywhayare23:09
gsomixsonney2k, 10 UTC.23:10
@sonney2kgsomix, ok then good luck for that :)23:10
@sonney2kgsomix, do you have time somewhen afterwards to chat?23:11
gsomixof course. e.g. tomorrow23:11
@sonney2kwhen?23:12
gsomixafter 15UTC, is this convenient for you?23:13
@sonney2kI only have time before that or later in the evening23:14
gsomixsonney2k, ok, than later in the evening. I'll be available.23:15
-!- heiko [~heiko@host86-177-176-121.range86-177.btcentralplus.com] has joined #shogun23:31
CIA-18shogun: Soeren Sonnenburg master * r053bdde / src/shogun/lib/SGReferencedData.h : destroy lock on free and fix race condition - http://git.io/ZDWpZQ23:32
blackburnah destroy..23:33
blackburnyeah makes sense23:33
@sonney2kgsomix, please still send the weekly report...23:34
blackburnsonney2k: btw I would like to get rid of spinlocks23:34
@sonney2kblackburn, no!23:34
@sonney2kwhy that?23:34
gsomixsonney2k, aha23:34
blackburnthey are unsafe - 100% cpu load for ever in case of glitch23:34
@sonney2kwhy unsafe?23:34
@sonney2kpthread mutexes have huge overhead23:35
@sonney2kin particular for these *fast* refcount inc/decrefs23:35
@sonney2kspinlocks are much much better suited23:35
shogun-buildbotbuild #31 of deb3 - modular_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb3%20-%20modular_interfaces/builds/31  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>23:36
shogun-buildbotbuild #27 of deb1 - libshogun started, including [8bc0fdfbf8a21ffd83a772043a3bd187f31223ca, 053bdde484b5f277de9acc7482931ee2fecb25e6]23:36
blackburnsonney2k: we don't need fast in sgobject in particular23:37
blackburnI'll become mad soon - when will it work finally?23:37
@sonney2kwell we use it to compute combined kernels etc23:37
blackburnsonney2k: any idea why it hangs?23:38
@sonney2kblackburn, relax and do sth else maybe - doc stuff?23:38
shogun-buildbotbuild #27 of deb1 - libshogun is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/2723:38
shogun-buildbotbuild #35 of deb2 - static_interfaces started, including [f1a01b70ec1519a0d586f5588d55b9927df84fc7]23:38
blackburnsonney2k: I have gsoc project here you know :D23:38
@sonney2kblackburn, ahum that is why you do SGSparse :D23:39
blackburnsonney2k: I do not really remember why did I do that23:39
blackburnahhh chris pushed me in these swarms full of monsters and dragons23:39
@sonney2kI guess because you promised at some point and then because you merged the broken commit from chris?23:39
blackburnoh you are so positive :D23:41
@sonney2kblackburn, I finally know why R fails23:42
blackburnwhy?23:42
@sonney2kit requires R >= 2.1323:42
@sonney2kand on bb is 2.1123:42
@sonney2kit would have worked with 2.11 if we had the old .r-install.sh script though23:42
blackburnno idea about it23:43
@sonney2kwell R 2.11 is from 201023:43
@sonney2kand they do such nice things like renaming .saveRDS to saveRDS23:43
blackburnsonney2k: I have seen R code a few times23:45
blackburnand I don't want to anymore23:46
blackburn:D23:46
@sonney2kI finally understand why it worked all good here23:46
@sonney2kR 2.15 ...23:46
@sonney2kI think we need to support the old version again :/23:49
gsomixgood night guys23:53
n4nd0good night gsomix23:56
-!- heiko [~heiko@host86-177-176-121.range86-177.btcentralplus.com] has left #shogun []23:57
@sonney2knite23:58
shogun-buildbotbuild #35 of deb2 - static_interfaces is complete: Failure [failed test]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb2%20-%20static_interfaces/builds/35  blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com>23:59
shogun-buildbotbuild #36 of deb2 - static_interfaces started, including [8bc0fdfbf8a21ffd83a772043a3bd187f31223ca, 053bdde484b5f277de9acc7482931ee2fecb25e6]23:59
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Quit: Page closed]23:59
--- Log closed Tue Jul 03 00:00:17 2012