--- Log opened Tue May 07 00:00:34 2013 | ||
pickle27 | this is what I get now: http://pastebin.com/LHqAyAUv | 01:30 |
---|---|---|
-!- sijin [~smuxi@144.214.222.109] has quit [] | 02:35 | |
-!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has quit [Quit: Leaving] | 02:58 | |
-!- gsomix [~gsomix@83.234.54.230] has quit [Ping timeout: 256 seconds] | 04:59 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 05:40 | |
-!- gsomix [~gsomix@83.234.54.230] has joined #shogun | 05:46 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 05:59 | |
-!- nube [~rho@116.90.239.3] has quit [Read error: Connection reset by peer] | 07:38 | |
-!- nube1 [~rho@116.90.239.3] has joined #shogun | 07:38 | |
-!- lisitsyn1 [~blackburn@83.234.54.230] has quit [Quit: Leaving.] | 08:31 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 08:36 | |
-!- erlenda [~erlenda@fw-oslo.intelcom.no] has joined #shogun | 08:47 | |
-!- nube1 [~rho@116.90.239.3] has quit [Quit: Leaving.] | 09:17 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 09:18 | |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has joined #shogun | 09:18 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 09:18 | |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 276 seconds] | 09:33 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 09:34 | |
-!- gsomix [~gsomix@83.234.54.230] has quit [Ping timeout: 272 seconds] | 09:41 | |
-!- gsomix [~gsomix@83.234.54.230] has joined #shogun | 09:41 | |
-!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 10:02 | |
-!- nube [~rho@116.90.239.3] has quit [Ping timeout: 245 seconds] | 10:35 | |
-!- nube [~rho@116.90.239.3] has joined #shogun | 11:03 | |
-!- sumit_ [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 11:37 | |
-!- van51 [~van51@athedsl-318776.home.otenet.gr] has joined #shogun | 11:49 | |
@wiking | sonney2k: yo | 11:52 |
@sonney2k | wiking, what? | 12:33 |
@wiking | sonney2k: ok so i didn't get it the last day | 12:34 |
-!- gsomix [~gsomix@83.234.54.230] has quit [Ping timeout: 272 seconds] | 12:34 | |
@wiking | sonney2k: about unit testing random | 12:34 |
@wiking | sonney2k: can i have a test in unit tests where i read a data from data submodule? | 12:34 |
-!- gsomix [~gsomix@83.234.54.230] has joined #shogun | 12:34 | |
@wiking | sonney2k: or that should rather be the part of integration test? | 12:34 |
@sonney2k | wiking, don't you think generating say 10 numbers and keeping them in ascii as unit test is sufficient? | 12:37 |
@sonney2k | or even generate 2x 10000 numbers with same seed | 12:37 |
@sonney2k | compare that they are the same | 12:37 |
@sonney2k | and check if the 10000th number is the same | 12:37 |
@wiking | sonney2k: ah nono | 12:40 |
@wiking | sonney2k: i mean what i do now is this | 12:40 |
@wiking | http://pastebin.com/xL6S0Lpc | 12:40 |
@wiking | this will generate the same numbers | 12:40 |
@wiking | the only 'problem' is that this does not test the arch/os independency | 12:41 |
@wiking | as for sure these 2 vectors are going to be the same | 12:41 |
@wiking | but they are not testing whether they are the same when it's on a different machine ;P | 12:41 |
@wiking | i mean different os/arch | 12:41 |
@wiking | should be of course ... :P | 12:42 |
@wiking | as the implementation is the same :) | 12:42 |
@wiking | but then again we should have a test for this | 12:42 |
@wiking | but we can do that then in the integration test | 12:42 |
-!- gsomix [~gsomix@83.234.54.230] has quit [Ping timeout: 272 seconds] | 12:46 | |
@sonney2k | wiking, yeah but if you just store the 10000th number - in some const then you have your test | 12:53 |
@wiking | sonney2k: well that's another option as well | 12:56 |
-!- iglesiasg [c1934d18@gateway/web/freenode/ip.193.147.77.24] has quit [Quit: Page closed] | 13:07 | |
thoralf | Whatever I do, shogun never stops throwing "*** glibc detected ***", "double free or corruption (!prev)". | 13:27 |
thoralf | gdb tells me the pointer - is there a way to find out, where this pointer has been allocated? | 13:27 |
thoralf | Using valgrind, gdb, etc.? | 13:27 |
@wiking | thoralf: valgrind it | 13:33 |
thoralf | wiking: It's a heisenbug. ;) | 13:35 |
thoralf | wiking: Valgrind does not tell me - instead it says "Address 0xb092b68 is 0 bytes after a block of size 8,184 alloc'd" (in foreign code) | 13:36 |
@wiking | well | 13:37 |
@wiking | with gdb and valgrind you pretty much can find any such bug | 13:37 |
@wiking | ;) | 13:37 |
thoralf | wiking: This claim needs proof... ;) | 13:41 |
@sonney2k | thoralf, compile w/o optimizations enabled then you will see :D | 13:43 |
-!- van51 [~van51@athedsl-318776.home.otenet.gr] has quit [Quit: Leaving.] | 13:44 | |
thoralf | sonney2k: I already compiled with --disable-optimization --enable-debug -O0 -g, etc. ;) | 13:44 |
@sonney2k | thoralf, then you are running multithreaded Iguess | 13:44 |
@wiking | thoralf: this claim has been proven by u running an OS | 13:45 |
@wiking | ;) | 13:45 |
@wiking | that has mouse, keyboard, screen, gpu etc | 13:45 |
thoralf | sonney2k: Yes, looks like. | 13:47 |
thoralf | How to turn threading off? | 13:47 |
thoralf | Will mkl->parallel->set_num_threads(1) be enough? | 13:49 |
thoralf | Okay, seems so. But at lease it did not change errors. | 13:50 |
@sonney2k | thoralf, how/when are you getting these double frees? | 14:04 |
@sonney2k | with streaming features? | 14:04 |
thoralf | Yes. | 14:06 |
@sonney2k | argh | 14:06 |
@sonney2k | yes that's all multi threaded | 14:06 |
@sonney2k | there is one thread reading in data | 14:06 |
@sonney2k | trying to keep some buffer full | 14:06 |
@sonney2k | and another doing the learning | 14:07 |
thoralf | sonney2k: It seems that "anyone" frees the .features in sparsevectors. | 14:07 |
thoralf | sonney2k: Copying all vectors seems to fix the double-frees, but introduces bigger leaks... | 14:08 |
@sonney2k | thoralf, when we did the conversion from e.g. double* vec, int vec_len -> SGVector | 14:08 |
@sonney2k | we didn't manage to finish this for everything yet | 14:08 |
@sonney2k | IIRC SparseFeatures still uses no SGSparseMatrix | 14:09 |
@sonney2k | which it should | 14:09 |
thoralf | sonney2k: I've been filling SGSparseMatrix by myself - with vectors from get_vector() ;) | 14:10 |
@sonney2k | no wonder then | 14:10 |
thoralf | sonney2k: So what's the intended use of the sparse vectors from get_vectors()? | 14:13 |
thoralf | Am I supposed to copy them? | 14:13 |
thoralf | copy as-in deep copy | 14:13 |
@sonney2k | thoralf, *intended* is that you create / get a sparse vector and use it / pass it around | 14:17 |
@sonney2k | freeing is done automagically | 14:17 |
thoralf | sonney2k: Okay, that's what I did. | 14:18 |
-!- nube [~rho@116.90.239.3] has quit [Quit: Leaving.] | 14:18 | |
thoralf | Except that freeing is done automagically by someone else. ;) | 14:18 |
@sonney2k | thoralf, yeah streaming features I guess | 14:19 |
thoralf | sonney2k: Yes, maybe it's related to this example buffer? | 14:19 |
thoralf | sonney2k: I initialized the example buffer to 1024 examples - are they freed or reused? | 14:20 |
@sonney2k | thoralf, well each example is a separately allocated | 14:21 |
@sonney2k | so it is just a list with 1k examples | 14:21 |
@sonney2k | and for sure they will be freed once you are done with them | 14:21 |
@sonney2k | thoralf, let me explain in detail how it is supposed to work once transitioned to the SG* structures: | 14:23 |
@sonney2k | the SGVector (or SGSparseVector etc) structures keep an internal reference count | 14:24 |
@sonney2k | so whenever you do SGVector x = y; | 14:24 |
@sonney2k | and y is some SGVector too | 14:24 |
@sonney2k | then we only copy references and size of the vector | 14:24 |
@sonney2k | to decrease refcount one needs to do | 14:25 |
@sonney2k | x=SGVector() | 14:25 |
@sonney2k | which will then assign an empty vector to x | 14:25 |
@sonney2k | decreasing the refcount for y again | 14:25 |
@sonney2k | and if y is no longer used | 14:26 |
@sonney2k | it will be destroyed too | 14:26 |
@sonney2k | now I guess you are seeing things like someone manually SG_FREE'ing the vector | 14:27 |
@sonney2k | and then it will later be destroyed -> boom | 14:27 |
thoralf | sonney2k: Yeah, thats what I'm thinking. | 14:27 |
@sonney2k | the correct fix is difficult :/ | 14:28 |
@sonney2k | maybe you can put NULL to the sgvector after the SG_FREE | 14:28 |
@sonney2k | as workaround | 14:29 |
@sonney2k | but you might suffer from an even more subtle situation: | 14:29 |
@sonney2k | consider you want to have an array of SGVectors | 14:30 |
@sonney2k | SGVector x[17]; | 14:30 |
@sonney2k | that is all good | 14:30 |
@sonney2k | but SGVector* x = malloc (...); | 14:31 |
@sonney2k | you get a kaboom | 14:31 |
@sonney2k | b'cause it needs the inplace new[]() - otherwise the constructor of sgvector is not called | 14:31 |
@sonney2k | and some stuff like refcount is not initialized | 14:31 |
@sonney2k | I tried to make this automatic by using SG_MALLOC - and for that malloc for SGVector is overloaded to call new[] | 14:32 |
@sonney2k | it is not ideal but I don't see a better way | 14:33 |
lisitsyn | sonney2k: it doesn't look normal to have SGVector* | 14:40 |
thoralf | sonney2k: shogun/io/streaming/ParseBuffer.h, finalize_example(), SG_FREE is called on ex_ring[i].fv.vector; the same fv.vector, has been passed to InputParser.h, which passes CStreamingSparseFeatures<T>::get_next_example() and current_vector. | 14:42 |
thoralf | sonney2k: The same reference then gets into SGSparseVector... q.e.d. ;) | 14:42 |
@sonney2k | thoralf, yeah | 14:42 |
@sonney2k | thoralf, if we would rewrite this to use | 14:43 |
thoralf | sonney2k: But I don't know how to fix ;) | 14:43 |
@sonney2k | some SGSpareseVector buf[1024] | 14:43 |
@sonney2k | then no more free is needed | 14:43 |
@sonney2k | but just a buf[10]=SGSparseVector() | 14:43 |
@sonney2k | to decrease the refcount of vector 10 | 14:44 |
thoralf | sonney2k: current_vector is SGSparseVectorEntry<T>* | 14:44 |
@sonney2k | that has to be refactored | 14:44 |
@sonney2k | it needs to become | 14:44 |
@sonney2k | SGSparseVector | 14:44 |
thoralf | sonney2k: in fact, current_vector is redundant, since it's the same as current_sgvector.features (only after calling get_vector()) | 14:46 |
@sonney2k | sounds like | 14:46 |
-!- sijin [~smuxi@144.214.222.109] has joined #shogun | 14:49 | |
thoralf | sonney2k: But in any case we need to copy the contents of current_vector, since ParseBuffer seems to take care of it (including deallocation). | 14:51 |
@sonney2k | thoralf, doesn't hurt - comparing this copy to a read from disk is like 1000 times faster | 14:51 |
thoralf | sonney2k: where does the disk read come from? | 14:52 |
@sonney2k | thoralf, streaming features! | 14:53 |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 15:17 | |
-!- Zheng [80b4f1e0@gateway/web/freenode/ip.128.180.241.224] has joined #shogun | 15:30 | |
Zheng | Hi, has anyone tried to handle this issue: https://github.com/shogun-toolbox/shogun/issues/1060? | 15:39 |
Zheng | An entrance issue | 15:40 |
-!- nube [~rho@49.244.79.215] has joined #shogun | 15:43 | |
-!- pickle27 [~kevin@rcv3-lab-pc.ee.queensu.ca] has joined #shogun | 15:45 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 15:49 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 16:05 | |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 16:14 | |
-!- sumit [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [Ping timeout: 245 seconds] | 16:39 | |
-!- Zheng [80b4f1e0@gateway/web/freenode/ip.128.180.241.224] has quit [Quit: Page closed] | 16:53 | |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has joined #shogun | 16:57 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 17:02 | |
-!- van51 [~van51@athedsl-318776.home.otenet.gr] has joined #shogun | 17:10 | |
-!- van51 [~van51@athedsl-318776.home.otenet.gr] has quit [Client Quit] | 17:11 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 17:25 | |
-!- gsomix [~gsomix@83.234.54.230] has joined #shogun | 17:45 | |
gsomix | hi | 17:45 |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:49 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has joined #shogun | 18:00 | |
-!- gsomix [~gsomix@83.234.54.230] has quit [Ping timeout: 255 seconds] | 18:15 | |
-!- nube [~rho@49.244.79.215] has quit [Ping timeout: 260 seconds] | 18:29 | |
-!- lisitsyn [~blackburn@188.168.3.6] has joined #shogun | 18:48 | |
-!- mode/#shogun [+o lisitsyn] by ChanServ | 18:48 | |
-!- vgorbati [d4029f22@gateway/web/freenode/ip.212.2.159.34] has quit [Ping timeout: 245 seconds] | 18:49 | |
-!- nube [~rho@49.244.21.43] has joined #shogun | 18:52 | |
-!- iglesiasg [d58f3201@gateway/web/freenode/ip.213.143.50.1] has joined #shogun | 18:55 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 18:56 | |
pickle27 | iglesiasg: hey I re-ran after building with no optimization and got this http://pastebin.com/LHqAyAUv | 18:57 |
@iglesiasg | pickle27: hey | 18:57 |
@iglesiasg | I checked the log you pasted last night | 18:57 |
@iglesiasg | let me see this one | 18:57 |
@iglesiasg | pickle27: but it is still the same error I am afraid | 18:58 |
@iglesiasg | vex amd64->IR: unhandled instruction bytes: 0xC5 0xF9 0x57 0xC0 0xE8 0x73 0x4 0xFF | 18:58 |
@iglesiasg | pickle27: isn't it? | 18:58 |
pickle27 | iglesiasg: yeah I saw that there in the middle but there was more upfront I thought might help | 18:58 |
@iglesiasg | pickle27: not really honestly | 18:59 |
pickle27 | iglesiasg: can we re-try the build server and see if it crashes when running the example again | 18:59 |
@iglesiasg | pickle27: I can try to take a look at it in my computer | 18:59 |
pickle27 | iglesiasg: okay. because I don't know what else I can do to debug if I can't replicate the actual crash on my pc | 18:59 |
@iglesiasg | pickle27: both the MLDA code and the example has been merged to upstream develop right? | 19:01 |
pickle27 | iglesiasg: yes it was merged yesterday | 19:01 |
@iglesiasg | pickle27: so I will take a look ASAP and give you and answer in the mailing list | 19:02 |
pickle27 | okay sounds good, thanks! | 19:02 |
@iglesiasg | probably by tomorrow midday | 19:03 |
@iglesiasg | so that is like 15h from now for me | 19:03 |
@iglesiasg | pickle27: I think we are in different time zones :) | 19:03 |
pickle27 | iglesiasg: yeah I've come to notice haha, Europe? | 19:04 |
@iglesiasg | pickle27: yep | 19:04 |
pickle27 | iglesiasg: cool, I'm in Canada | 19:04 |
@iglesiasg | pickle27: nice! I will try to take more care of my English then :P | 19:06 |
@iglesiasg | well are you in English or French speaking area? | 19:06 |
pickle27 | iglesiasg: I'm in an English area, I do speak a bit of french though, whereabouts are you in Europe? | 19:06 |
@iglesiasg | pickle27: right now in Spain, I live in Sweden though | 19:07 |
pickle27 | iglesiasg: neat! | 19:07 |
thoralf | pickle27: Tried to add --enable-debug --disable-optimization to ./configure? This should remove the illegal instructions in valgrind. | 19:24 |
@iglesiasg | thoralf: I think he has already tried compiling with the disabel-optimization option | 19:27 |
thoralf | iglesiasg: Maybe, but as he didn't mention it in his post, I thought it could save him some trial and error. :) | 19:31 |
@iglesiasg | sonney2k suggested doing that yesterday | 19:32 |
thoralf | iglesiasg: It still looks too familiar to me... ;) | 19:36 |
thoralf | pickle27: We're wondering if you tried (in shogun/src): make clean; ./configure --enable-debug --disable-optimization | 19:37 |
@iglesiasg | :) | 19:37 |
votjakovr | Hi guys :) Has anyone run integration tests on local machine? Is regression_gaussian_process_modular.py failed? | 19:49 |
@sonney2k | votjakovr, seems it works: http://shogun-toolbox.org/buildbot/waterfall | 19:52 |
@sonney2k | why? | 19:53 |
@sonney2k | wiking, ubu1 - libshogun buildbot is down | 19:53 |
thoralf | sonney2k: Cool visualization. | 19:54 |
votjakovr | sonney2k: Oh, but travis-ci shows an errors. | 19:54 |
@sonney2k | thoralf, what? | 19:56 |
thoralf | sonney2k: The buildbot/waterfall | 19:57 |
@sonney2k | thoralf, ahh ok | 19:58 |
@sonney2k | wasn't so impressive to me.. | 19:58 |
@sonney2k | votjakovr, checking | 19:59 |
@sonney2k | votjakovr, hmmhh the returned label vector is *slightly* different | 19:59 |
@sonney2k | any idea why that could be? | 19:59 |
@sonney2k | any randomness in there? | 19:59 |
@sonney2k | or different numerical libraries? | 20:00 |
@sonney2k | iglesiasg, don't you have ubuntu? | 20:05 |
@iglesiasg | sonney2k: I currently use more arch than ubuntu | 20:06 |
@iglesiasg | sonney2k: why? | 20:06 |
@sonney2k | I wish we could reproduce the travis failure | 20:06 |
@sonney2k | travis runs on ubuntu | 20:06 |
votjakovr | sonney2k: Nope, when I changed implementation of ExactInferenceMethod, I regenerated test data. And everything is OK on my machine. | 20:07 |
@iglesiasg | lisitsyn uses ubuntu I think | 20:07 |
votjakovr | sonney2k: And I use Ubuntu 13.04 | 20:07 |
@sonney2k | votjakovr, which version of numpy? | 20:07 |
@sonney2k | 1.6.1 or .2? | 20:07 |
pickle27 | thoralf: I will try now | 20:07 |
votjakovr | sonney2k: 1.7.1 | 20:08 |
@sonney2k | votjakovr, hmm - labels is generated randomly though | 20:11 |
@sonney2k | votjakovr, I just checked the diff | 20:15 |
@sonney2k | for some reason the 100 dim label vector differs in 3 entries | 20:15 |
@sonney2k | 1.144566753132682 vs 1.144566753132681 | 20:15 |
@sonney2k | 1.15097403962122 vs 1.150974039621219 | 20:15 |
@sonney2k | 1.447796756701278 vs 1.447796756701277 | 20:16 |
-!- madhura [~madhura@59.95.237.64] has joined #shogun | 20:16 | |
@sonney2k | wiking, is there any way to specify the numpy version in travis? | 20:18 |
votjakovr | sonney2k: Yep, I have done the same. But I have no idea why they are different... | 20:18 |
votjakovr | sonney2k: Maybe some problem with floating point precision | 20:19 |
@sonney2k | votjakovr, me neither and I don't understand how the rest of the output can match perfectly | 20:19 |
@sonney2k | votjakovr, does the algorithm use threads or anything | 20:28 |
@lisitsyn | what's up? | 20:32 |
-!- gsomix [~gsomix@188.168.3.6] has joined #shogun | 20:32 | |
votjakovr | sonney2k: there's just operations with matrices | 20:33 |
@lisitsyn | sonney2k: have you seen? we've got some slots | 20:34 |
@sonney2k | lisitsyn, no I haven't | 20:34 |
@lisitsyn | sonney2k: 8 | 20:34 |
naywhayare | nice! I was hoping for that many but just like I was told to expect we got only two :) | 20:35 |
@sonney2k | cool | 20:35 |
gsomix | lisitsyn, sonney2k hey, guys, wuts goin' on? | 20:35 |
@sonney2k | so we only have to unaccept 2 :D | 20:35 |
@lisitsyn | hah yes currently | 20:35 |
@sonney2k | we had 9 last year but used only 8 right? | 20:36 |
@lisitsyn | yes | 20:36 |
@sonney2k | naywhayare, be grateful it's a lot of work | 20:38 |
@iglesiasg | lisitsyn: oh we already got an answer? | 20:38 |
@lisitsyn | iglesiasg: yeah carols is fast this time | 20:38 |
@iglesiasg | I thought it was going to be tom! | 20:38 |
@lisitsyn | I guess it may be changed | 20:38 |
@sonney2k | she wants to be on vacations soon so :) | 20:38 |
@iglesiasg | hehe | 20:39 |
@lisitsyn | oh! I should ask her about delivering t-shirt to you sonney2k | 20:39 |
@lisitsyn | are you ok with it? | 20:39 |
@iglesiasg | 8 is a nice number anyway | 20:39 |
@iglesiasg | let's do some vodking | 20:39 |
@lisitsyn | iglesiasg: you know the way | 20:39 |
@iglesiasg | sonney2k: is it a good idea to tweet about this? | 20:39 |
@sonney2k | why not | 20:40 |
naywhayare | sonney2k: yeah, you are probably right... | 20:40 |
naywhayare | (well... more than probably. you've done this before :)) | 20:40 |
@lisitsyn | twice | 20:40 |
@lisitsyn | :D | 20:40 |
@sonney2k | naywhayare, it really depends on your students/mentors though. with 8 it already is quite a bit of coordination. and if one student/mentor sucks you can easily cope with that | 20:41 |
@sonney2k | when you have 'just' 2 slots | 20:41 |
naywhayare | yeah; we got a number of good applications though, so it will be tough to choose which two to accept | 20:42 |
naywhayare | even so, with how busy I tend to be, it's probably a good thing in terms of "work/life balance" (not that I get to know what that is as a grad student...) | 20:42 |
@sonney2k | naywhayare, what we did was ask our top candidates all for patches | 20:42 |
@sonney2k | even already for their project | 20:42 |
@sonney2k | then you select the most engaged | 20:43 |
@sonney2k | and that obviously wasn't the wrong decision | 20:43 |
naywhayare | that is probably a good idea; a few of our candidates have already taken the initiative and started doing that | 20:43 |
@sonney2k | lisitsyn and heiko are still around :D | 20:43 |
@sonney2k | 3 years later now | 20:43 |
@lisitsyn | yeah I am a gsoc child | 20:43 |
@lisitsyn | :D | 20:43 |
@lisitsyn | no, I am here for two years I think | 20:44 |
@sonney2k | lisitsyn, true | 20:45 |
@sonney2k | 2011 | 20:45 |
@sonney2k | heiko dates back to 2010 | 20:46 |
@wiking | 1234 | 20:55 |
@lisitsyn | wiking: good! | 20:56 |
@wiking | yes keyboard works! | 20:56 |
@wiking | :> | 20:56 |
-!- hoijui [~hoijui@dslb-088-074-109-050.pools.arcor-ip.net] has joined #shogun | 20:56 | |
@sonney2k | wiking, get the ubuntu buildbot back to work! | 20:56 |
@wiking | oh | 20:56 |
@wiking | sonney2k: i would if i would access to that machine | 20:56 |
@wiking | it's down and i really have no clue when it's going to be back | 20:56 |
@sonney2k | wiking, hmm!? | 20:57 |
@wiking | the machine | 20:57 |
@wiking | it's down | 20:57 |
@wiking | or somebody took out the utp cable | 20:57 |
@wiking | or i have no clue | 20:57 |
@wiking | anyhow i have no physical access to that machine | 20:57 |
@wiking | so i cannot do anything about it | 20:57 |
@sonney2k | wiking, ohh and could you merge from develop -> master the configure change for swig2.0 or what do we do about master failing? | 20:57 |
@sonney2k | on bsd1? | 20:57 |
@wiking | mmmm | 20:57 |
@wiking | noclue :) | 20:58 |
@wiking | i'll check tonight | 20:58 |
-!- nube [~rho@49.244.21.43] has quit [Ping timeout: 264 seconds] | 20:58 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 20:58 | |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * b466c0d / src/shogun/lib/tapkee/ (11 files): https://github.com/shogun-toolbox/shogun/commit/b466c0dcca4467d7c219619367488b104fc88e1b | 20:58 |
shogun-notifier- | shogun: Updated tapkee library | 20:58 |
shogun-notifier- | shogun: | 20:58 |
shogun-notifier- | shogun: - Proper non-negativity checks in parameters | 20:58 |
shogun-notifier- | shogun: - VP-tree neighbors method | 20:58 |
shogun-notifier- | shogun: - Some minor fixes | 20:58 |
-!- nube [~rho@49.244.128.165] has joined #shogun | 21:01 | |
shogun-buildbot | build #1110 of deb1 - libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1110 blamelist: Sergey Lisitsyn <lisitsyn.s.o@gmail.com> | 21:03 |
@lisitsyn | oops! | 21:03 |
@lisitsyn | sorry | 21:03 |
@sonney2k | guys punish him! | 21:03 |
@lisitsyn | fixing | 21:04 |
@lisitsyn | my script for auto-conversion of includes failed here | 21:04 |
@iglesiasg | hehe shit happens | 21:05 |
@lisitsyn | iglesiasg: bad you are not in sweden, ice hockey event there | 21:06 |
@lisitsyn | rus-usa right now | 21:06 |
@sonney2k | lisitsyn, ohh cold war! | 21:07 |
-!- travis-ci [~travis-ci@ec2-54-234-229-115.compute-1.amazonaws.com] has joined #shogun | 21:07 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn'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/6966751 | 21:07 |
-!- travis-ci [~travis-ci@ec2-54-234-229-115.compute-1.amazonaws.com] has left #shogun [] | 21:07 | |
* sonney2k sigh | 21:07 | |
@sonney2k | we have to kick out 2 students now | 21:07 |
@sonney2k | if anyone has a better idea than throwing a dice... say so | 21:08 |
thoralf | sonney2k: Random numbers needed? http://xkcd.com/221/ | 21:09 |
thoralf | (off-topic) | 21:10 |
@sonney2k | thoralf, I choose the seed :P | 21:10 |
@sonney2k | and wiking provides the prng | 21:10 |
shogun-notifier- | shogun: Sergey Lisitsyn :develop * 542609f / src/shogun/lib/tapkee/neighbors/vptree.hpp: https://github.com/shogun-toolbox/shogun/commit/542609f99d210296a9a9fed81a657914548b29af | 21:11 |
shogun-notifier- | shogun: Fixed include in vptree | 21:11 |
-!- Zheng [80b422fb@gateway/web/freenode/ip.128.180.34.251] has joined #shogun | 21:14 | |
shogun-buildbot | build #1111 of deb1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/deb1%20-%20libshogun/builds/1111 | 21:14 |
@wiking | sonney2k: why so? | 21:20 |
@wiking | sonney2k: u think we wont' get enough slots? | 21:20 |
@sonney2k | wiking, we got 8 | 21:20 |
@wiking | aaah it's already decided? | 21:20 |
@wiking | i mean the slots? | 21:20 |
@sonney2k | living behind the moon aren't ya? | 21:20 |
@wiking | i missed that | 21:21 |
@wiking | sorry man | 21:21 |
@wiking | too busy in the last 1 month | 21:21 |
@sonney2k | wiking, yeah we have to kick out 2 students | 21:21 |
@wiking | mmmm | 21:21 |
@wiking | and all those 10 are tops? | 21:21 |
@wiking | i mean super great applications? | 21:21 |
@sonney2k | yes | 21:21 |
shogun-buildbot | build #925 of bsd1 - libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/bsd1%20-%20libshogun/builds/925 | 21:21 |
@wiking | mmmm | 21:21 |
@wiking | well if u cannot really prioritize | 21:21 |
@wiking | then let's kick out 3 and 7 | 21:22 |
@wiking | :) | 21:22 |
@wiking | i mean i was just your random machine | 21:22 |
@sonney2k | no prime numbers have to stay :D | 21:22 |
@wiking | and i'm totally objective as i dont even know | 21:22 |
thoralf | Good choice. ;) | 21:22 |
@wiking | the applications | 21:22 |
@wiking | sonney2k: hehehe well u can assign any of the applications to be 3 and 7 | 21:22 |
@wiking | :P | 21:22 |
@wiking | haven't defined the order of the applications | 21:22 |
@wiking | but then again | 21:23 |
@wiking | other option is a very stupid one | 21:23 |
@wiking | order them by submission date | 21:23 |
@wiking | and remove the last 2 | 21:23 |
@wiking | i mean it's really stupid | 21:23 |
@wiking | but it's another option | 21:23 |
@wiking | oooooor | 21:23 |
@wiking | let's start gathering money for those 2 | 21:24 |
@wiking | \o/ | 21:24 |
@wiking | go to kaggle | 21:24 |
@wiking | win a competition | 21:24 |
@wiking | and then we can support them | 21:24 |
@wiking | :D | 21:24 |
@wiking | sonney2k: or give me the list and i'll review them for ya ;) | 21:24 |
@wiking | i bet i can find you 2 not so good ones :DDD | 21:24 |
@wiking | even 3 :P | 21:24 |
* wiking is accepting bribes :DDD | 21:25 | |
@wiking | sonney2k: or why not to somehow prio them by the entrance | 21:25 |
@wiking | PRs | 21:25 |
@wiking | i bet we can measure there a very nice objective way | 21:26 |
@wiking | ok | 21:26 |
@wiking | retweeted | 21:26 |
@wiking | btw who's not here out of those 10 ppl? :) | 21:27 |
@sonney2k | wiking, a couple :D | 21:27 |
@wiking | heh | 21:27 |
-!- Zheng [80b422fb@gateway/web/freenode/ip.128.180.34.251] has quit [Ping timeout: 245 seconds] | 21:27 | |
@wiking | theeere you gooooo | 21:27 |
@wiking | :)))) | 21:27 |
@wiking | no IRC no GSOC | 21:27 |
@wiking | :DDD | 21:27 |
@wiking | pretty standard measure :P | 21:28 |
-!- iglesiasg [d58f3201@gateway/web/freenode/ip.213.143.50.1] has quit [Ping timeout: 245 seconds] | 21:30 | |
pickle27 | haha good thing I'm on IRC right now lol | 21:30 |
-!- travis-ci [~travis-ci@ec2-54-234-229-115.compute-1.amazonaws.com] has joined #shogun | 21:35 | |
travis-ci | [travis-ci] it's Sergey Lisitsyn'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/6967091 | 21:35 |
-!- travis-ci [~travis-ci@ec2-54-234-229-115.compute-1.amazonaws.com] has left #shogun [] | 21:35 | |
-!- Zheng [80b422fb@gateway/web/freenode/ip.128.180.34.251] has joined #shogun | 21:38 | |
-!- thoralf [~thoralf@enki.zib.de] has left #shogun ["Konversation terminated!"] | 21:39 | |
-!- hoijui [~hoijui@dslb-088-074-109-050.pools.arcor-ip.net] has quit [Quit: Leaving] | 21:41 | |
pickle27 | the people involved in this left but I just re-ran the valgrind after compiling shogun with debugging symbols and got the same result, http://pastebin.com/00rU2vcU | 21:45 |
pickle27 | iglesias is going to try on his machine and see if he can reproduce the error because I can not - the classifier_lda example still runs fine on my system | 21:47 |
@sonney2k | pickle27, what is the issue? | 22:06 |
gsomix | wiking, hey, you're awesome. | 22:10 |
-!- Zheng [80b422fb@gateway/web/freenode/ip.128.180.34.251] has quit [Quit: Page closed] | 22:10 | |
@sonney2k | hmmhh I would wish we could have sprint meetings like scikits https://github.com/scikit-learn/scikit-learn/wiki/Past-sprints | 22:30 |
@sonney2k | like meet once / year in berlin or so | 22:30 |
@sonney2k | and then do a week of coding vodking :D | 22:30 |
-!- poojits [75d35d4a@gateway/web/freenode/ip.117.211.93.74] has joined #shogun | 23:03 | |
@wiking | gsomix: :D | 23:05 |
-!- votjakovr [~votjakovr@host-46-241-3-209.bbcustomer.zsttk.net] has quit [Quit: Leaving] | 23:13 | |
-!- mona2302 [4e080b1f@gateway/web/freenode/ip.78.8.11.31] has joined #shogun | 23:15 | |
-!- mona2302 [4e080b1f@gateway/web/freenode/ip.78.8.11.31] has quit [Quit: Page closed] | 23:24 | |
-!- iglesiasg [d58f3213@gateway/web/freenode/ip.213.143.50.19] has joined #shogun | 23:36 | |
iglesiasg | pickle27: hey, I just checked logs | 23:38 |
iglesiasg | pickle27: for some reason the last pastebin link doesn't open find | 23:39 |
-!- Ingvaras [522c194e@gateway/web/freenode/ip.82.44.25.78] has joined #shogun | 23:39 | |
iglesiasg | oh fine* | 23:39 |
iglesiasg | pickle27: anyway, if valgrind complains about it, it is a sign that it doesn't run fine in your system ;) | 23:39 |
--- Log closed Wed May 08 00:00:35 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!