| --- Log opened Mon Jul 09 00:00:17 2012 | ||
| CIA-18 | shogun: Sergey Lisitsyn openmp * r4e109e3 / (2 files): Made CDistance use openmp - http://git.io/l24RNA | 00:03 |
|---|---|---|
| CIA-18 | shogun: Sergey Lisitsyn openmp * r2b811d5 / (2 files): Made CombinedKernel use openmp - http://git.io/DnBP-g | 00:25 |
| CIA-18 | shogun: Sergey Lisitsyn openmp * r15203a3 / src/shogun/lib/external/shogun_libsvm.cpp : Made LibSVM use openmp - http://git.io/pZMErA | 00:37 |
| CIA-18 | shogun: Sergey Lisitsyn openmp * r6cc5b15 / (2 files): Made DistanceMachine use openmp - http://git.io/_EN9_g | 00:37 |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 00:47 | |
| -!- blackburn [~blackburn@81.28.187.0] has quit [Quit: Leaving.] | 01:47 | |
| shogun-buildbot | build #18 of nightly_default is complete: Failure [failed updatedoc] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/18 | 03:36 |
| -!- av3ngr [av3ngr@nat/redhat/x-xlpzesjskamqquhz] has joined #shogun | 04:11 | |
| -!- av3ngr [av3ngr@nat/redhat/x-xlpzesjskamqquhz] has left #shogun [] | 04:12 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 09:38 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 244 seconds] | 09:46 | |
| -!- uricamic [~uricamic@2001:718:2:1634:54ea:8a06:2501:ed7b] has joined #shogun | 09:50 | |
| -!- gsomix [~gsomix@178.45.92.193] has joined #shogun | 09:56 | |
| gsomix | good morning | 09:56 |
| gsomix | sonney2k, got your mail. working on this problem now | 10:03 |
| -!- blackburn [~blackburn@81.28.187.0] has joined #shogun | 10:05 | |
| CIA-18 | shogun: Sergey Lisitsyn master * r47a4297 / src/shogun/mathematics/lapack.cpp : Made lapack use mallocs instead of SG_MALLOCs - http://git.io/VZFXZA | 10:44 |
| -!- blackburn [~blackburn@81.28.187.0] has quit [Quit: Leaving.] | 10:44 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 10:49 | |
| -!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 10:57 | |
| -!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 10:57 | |
| @sonney2k | gsomix, can you reproduce the problem? And do you think it is our fault? | 11:15 |
| gsomix | sonney2k, problems with trace mallocs are related to multithread. blackburn fixed already. but there is problem with CMap. | 11:17 |
| gsomix | and I think this is my fault. | 11:18 |
| -!- pluskid [~pluskid@li411-226.members.linode.com] has joined #shogun | 11:28 | |
| -!- pluskid [~pluskid@li411-226.members.linode.com] has quit [Ping timeout: 264 seconds] | 11:39 | |
| -!- pluskid [~pluskid@111.120.6.180] has joined #shogun | 11:52 | |
| -!- gsomix [~gsomix@178.45.92.193] has quit [Read error: Connection reset by peer] | 11:54 | |
| shogun-buildbot | build #19 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/19 | 11:55 |
| -!- gsomix [~gsomix@88.200.242.178] has joined #shogun | 12:10 | |
| wiking | boaaah | 12:56 |
| gsomix | sonney2k, around? it seems, that CMap works not properly with multithreaded insert/delete. | 13:18 |
| gsomix | see you later guys | 13:18 |
| -!- gsomix [~gsomix@88.200.242.178] has quit [Ping timeout: 246 seconds] | 13:22 | |
| -!- gsomix [~gsomix@88.200.242.178] has joined #shogun | 13:53 | |
| wiking | n4nd0: ping | 14:05 |
| n4nd0 | wiking: pang | 14:05 |
| wiking | n4nd0: pong ;) | 14:06 |
| wiking | ok so question: svm^struct is going to be in ? | 14:06 |
| n4nd0 | in like in shogun? | 14:06 |
| wiking | yes | 14:06 |
| n4nd0 | not exactly that implementation | 14:07 |
| n4nd0 | what I am doing is a similar project though | 14:07 |
| wiking | since currently u have mosek but i was wondering if u are planning to include svm^struct as well | 14:08 |
| n4nd0 | no, I am not planning that | 14:08 |
| n4nd0 | but we are planning to do something alternative that does not require mosek | 14:08 |
| wiking | ah good that's good enough | 14:08 |
| n4nd0 | nice | 14:10 |
| -!- pluskid [~pluskid@111.120.6.180] has quit [Quit: Leaving] | 14:36 | |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 244 seconds] | 15:12 | |
| -!- nietpiet [92322838@gateway/web/freenode/ip.146.50.40.56] has joined #shogun | 15:19 | |
| -!- emrecelikten [~emre@213.153.218.163] has joined #shogun | 15:45 | |
| -!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has joined #shogun | 17:15 | |
| -!- gsomix [~gsomix@88.200.242.178] has quit [Ping timeout: 246 seconds] | 17:35 | |
| -!- gsomix [~gsomix@88.200.242.178] has joined #shogun | 17:35 | |
| -!- uricamic [~uricamic@2001:718:2:1634:54ea:8a06:2501:ed7b] has quit [Quit: Leaving.] | 17:43 | |
| -!- blackburn [~blackburn@81.28.187.0] has joined #shogun | 18:07 | |
| blackburn | sonney2k: sg_malloc is not thread-safe | 18:07 |
| gsomix | and also my CMap | 18:11 |
| heiko | gsomix, do you have any idea about this trace-mallocs weirdness? | 18:13 |
| blackburn | heiko: thread-unsafetyy | 18:15 |
| -!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 18:15 | |
| -!- cwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has quit [Ping timeout: 245 seconds] | 18:20 | |
| heiko | blackburn, really? that is it? | 18:21 |
| blackburn | heiko: well we use cmap to store allocated blocks | 18:22 |
| blackburn | and insert is not thread safe | 18:22 |
| heiko | mmh | 18:22 |
| heiko | but that should be easy to fix | 18:22 |
| heiko | just lock whil inserting | 18:22 |
| blackburn | sure | 18:22 |
| heiko | blackburn, around? | 19:06 |
| blackburn | yeah | 19:06 |
| heiko | computing covariance matrices... | 19:06 |
| heiko | cblas_dgemm ?? | 19:06 |
| blackburn | sure | 19:06 |
| heiko | I saw that in your code | 19:07 |
| blackburn | what is the problem? | 19:07 |
| heiko | what method is that? | 19:07 |
| blackburn | dot product of matrices | 19:07 |
| heiko | ah alright | 19:07 |
| blackburn | well to be honest it is not covariance | 19:07 |
| blackburn | but once we center it | 19:07 |
| blackburn | or center features before | 19:07 |
| blackburn | it is | 19:07 |
| heiko | which way is best to compute cov(h) where h is a vector? | 19:08 |
| heiko | like the matlab equivalent | 19:08 |
| heiko | that way youre doing there? | 19:08 |
| blackburn | dger I guess | 19:08 |
| heiko | which is? | 19:08 |
| heiko | I am not used to all these libraries :) | 19:09 |
| blackburn | heiko: well what is cov(h)? | 19:09 |
| blackburn | x'*x resulting with matrix? | 19:09 |
| blackburn | if so cblas_dger does the job | 19:09 |
| heiko | empirical covariance matrix where the entries contains the covariance of pairs of the random variables in the input matrix | 19:11 |
| heiko | which would be E[hh^T] - E[h]E[h]^T | 19:11 |
| heiko | and expectations are estimated by sample mean | 19:11 |
| blackburn | heiko: I'd suggest to cblas_dger and center_matrix after | 19:11 |
| heiko | ok alright | 19:12 |
| heiko | I will add a routine to CStatistics | 19:12 |
| heiko | this blas stuff is horrible to read ;) | 19:12 |
| blackburn | no if you are android | 19:12 |
| heiko | what? | 19:12 |
| blackburn | I am | 19:12 |
| blackburn | :D | 19:12 |
| heiko | lol :D | 19:12 |
| heiko | an evil robot, that what you are! | 19:13 |
| heiko | what does dger do? | 19:13 |
| blackburn | heiko: just product of vectors | 19:13 |
| heiko | inner or outer? | 19:13 |
| heiko | result is matrix or vecor? | 19:13 |
| blackburn | hehe | 19:13 |
| blackburn | good shot | 19:13 |
| blackburn | matrix | 19:13 |
| heiko | ok so I can do the hh^T thing with it | 19:14 |
| blackburn | but matrix / scalar is the choice | 19:14 |
| heiko | nice | 19:14 |
| blackburn | heiko: yes but do not forget to zero resulting matrix before | 19:14 |
| heiko | oh man | 19:14 |
| heiko | complicated | 19:14 |
| heiko | will add that | 19:15 |
| blackburn | cblas_dger(n_dims,n_dims,1.0,x,1,x,1,cov_matrix,n_dims) | 19:15 |
| blackburn | heiko: ^ | 19:15 |
| blackburn | cov_matrix += xx' that is | 19:15 |
| blackburn | or may be | 19:15 |
| blackburn | cov_matrix += 1.0*xx' | 19:15 |
| heiko | mmh ok | 19:16 |
| heiko | Ill dig into it | 19:16 |
| blackburn | heiko: if you are confused with ld* - it is just increment of dimensions | 19:16 |
| blackburn | so 1 for vectors and #num_of_rows for matrix | 19:17 |
| heiko | argh | 19:24 |
| heiko | anoyying | 19:25 |
| blackburn | why? | 19:25 |
| heiko | because I still have no idea what to call ;) and I hate try and error code development | 19:28 |
| blackburn | heiko: what is wrong with dger? | 19:28 |
| heiko | its fine, I just hat these technical things, but Ill stop complaining now | 19:30 |
| blackburn | heiko: just ask me :) | 19:31 |
| heiko | well, if you know what to do, could you add a method static SGMatrix<T> cov(SGMatrix<T> observations) to CStatistics? :) | 19:31 |
| blackburn | heiko: sure | 19:32 |
| blackburn | heiko: observations is column-wise feature vectors? | 19:32 |
| heiko | it would be great if there was a bool flag for transposing | 19:32 |
| heiko | yes | 19:32 |
| blackburn | heiko: transposing what? observations? | 19:32 |
| heiko | yes I mean if my data is the wrong way round it would be nice if I wouldnt have to transpose it before | 19:32 |
| heiko | but its not that important | 19:33 |
| blackburn | hahha | 19:33 |
| blackburn | I just found | 19:33 |
| blackburn | pseudo-formulas | 19:33 |
| blackburn | in statistics | 19:33 |
| blackburn | nicee | 19:33 |
| heiko | what? | 19:34 |
| heiko | Ah I see :) | 19:34 |
| blackburn | loooool | 19:34 |
| blackburn | integral looks very lovely | 19:34 |
| heiko | hehe I might translate these into latex | 19:34 |
| heiko | taken from ALGLIB :) | 19:34 |
| blackburn | ok ETA 20 minutes | 19:34 |
| heiko | many thanks :) | 19:35 |
| heiko | blackburn, they look horrible in doxygen documentation ;) I will definately change that | 19:40 |
| -!- nietpiet [92322838@gateway/web/freenode/ip.146.50.40.56] has quit [Quit: Page closed] | 19:41 | |
| blackburn | heiko: can you test it afterwards? | 19:45 |
| heiko | I will :) | 19:45 |
| blackburn | hmm heiko it looks like in LLE I call kernel matrix as covariance matrix | 19:46 |
| heiko | ah thats even better, I can copy it | 19:46 |
| heiko | which lines? | 19:47 |
| heiko | blackburn, is it the one in 498? | 19:47 |
| blackburn | heiko: shogun/converter/LocallyLinearEmbedding.cpp:309 | 19:48 |
| heiko | thats dgemm | 19:48 |
| blackburn | heiko: yes | 19:48 |
| blackburn | X'X | 19:48 |
| blackburn | heiko: linear kernel matrix that is | 19:49 |
| @sonney2k | blackburn, well then we have to make it so | 19:50 |
| blackburn | sonney2k: anyway lapacks mallocs are out of scope for tracing | 19:50 |
| @sonney2k | going back to just malloc will not solve the problem | 19:50 |
| blackburn | sonney2k: no but it shouldn't be tracked | 19:50 |
| @sonney2k | blackburn, but there is no error check in there so that won't do it | 19:51 |
| @sonney2k | blackburn, so what is not thread safe about them - the tracing stuff only right? | 19:53 |
| gsomix | well time to work | 19:56 |
| gsomix | good evening | 19:56 |
| CIA-18 | shogun: Soeren Sonnenburg master * rda0a349 / src/shogun/mathematics/lapack.cpp : Revert "Made lapack use mallocs instead of SG_MALLOCs" - http://git.io/Anljog | 19:56 |
| @sonney2k | gsomix, hi there - regarding thread safety of CMap - I guess all you would need to do is change the add / remove functions right? | 20:00 |
| @sonney2k | gsomix, if not other option would be to have a general acquire_lock() / release_lock() in CMap that one always has to call before doing sth with CMap | 20:01 |
| heiko | blackburn, finally got it :) | 20:08 |
| @sonney2k | wiking, don't miss your progress report this week! | 20:12 |
| blackburn | heiko: sorry was away for a while | 20:20 |
| blackburn | will commit soon | 20:21 |
| heiko | blackburn | 20:21 |
| heiko | I just found that I already did this | 20:21 |
| heiko | SGMatrix has matrix multiply | 20:21 |
| blackburn | :D | 20:21 |
| -!- alexlovesdata [c25fae98@gateway/web/freenode/ip.194.95.174.152] has joined #shogun | 20:21 | |
| heiko | using cblas_dgemm | 20:21 |
| blackburn | heiko: okay should I do that then? | 20:21 |
| heiko | sorry :( | 20:21 |
| blackburn | no problem :D | 20:21 |
| heiko | just need to center matrix before then its fine | 20:21 |
| blackburn | heiko: after you mean? | 20:22 |
| heiko | no before | 20:22 |
| heiko | and its not centering as center() method | 20:22 |
| heiko | but only remove mean of each column | 20:22 |
| blackburn | yeah exactly what I mean | 20:22 |
| blackburn | heiko: it is the same | 20:22 |
| heiko | are you sure? | 20:23 |
| blackburn | pretty much | 20:23 |
| heiko | oh yeah | 20:23 |
| heiko | just seeing it | 20:23 |
| blackburn | but it is more efficient | 20:23 |
| heiko | man, i should sleep more :) | 20:23 |
| blackburn | and more | 20:23 |
| blackburn | it does not modify matrix | 20:24 |
| blackburn | feature one I mean | 20:24 |
| heiko | not modify? | 20:24 |
| heiko | center() you mean() | 20:24 |
| heiko | ehm | 20:24 |
| heiko | how to not modify matrix? | 20:24 |
| heiko | ah | 20:25 |
| heiko | actually | 20:25 |
| heiko | one could do this with the +beta*A thing at the end of dgem | 20:25 |
| blackburn | heiko: well you don't have to subtract mean and to modify feature matrix | 20:25 |
| blackburn | only covariance is to be modified | 20:25 |
| heiko | afterwards? | 20:26 |
| blackburn | yeah I think so | 20:26 |
| blackburn | let me check | 20:26 |
| heiko | are you sure? | 20:26 |
| heiko | no I think it has to be done before | 20:26 |
| blackburn | heiko: before what? | 20:26 |
| heiko | before computing product | 20:26 |
| heiko | I mean one has to center feature matrix | 20:27 |
| blackburn | heiko: argh sorry | 20:27 |
| blackburn | center works only with kernel matrix it seems | 20:27 |
| heiko | yeah there its fine | 20:28 |
| blackburn | heiko: just checked, yeah | 20:30 |
| blackburn | sad it doesn't work with covariance | 20:30 |
| @sonney2k | gsomix, why do you think that is slow? | 21:03 |
| gsomix | sonney2k, I had to do it at morning... :) | 21:06 |
| @sonney2k | gsomix, I mean basically none of the shogun stuff is thread safe | 21:06 |
| @sonney2k | we make things thread safe when we need it and I am not so sure about CMap here | 21:07 |
| blackburn | sonney2k: why to make it explicit | 21:07 |
| @sonney2k | blackburn, you mean the locking/ unlocking? | 21:08 |
| blackburn | yes | 21:08 |
| @sonney2k | if it significantly slows down the 'normal' use case then we should make it explicit | 21:08 |
| gsomix | sonney2k, aww, just got your mail | 21:09 |
| blackburn | sonney2k: slows down significantly?? how | 21:09 |
| blackburn | lock is not that slow | 21:09 |
| @sonney2k | blackburn, did you measure? | 21:10 |
| blackburn | sonney2k: did you? | 21:10 |
| blackburn | I didn't | 21:11 |
| @sonney2k | no | 21:11 |
| @sonney2k | blackburn, but basically no data structure or anything in shogun is thread safe | 21:11 |
| blackburn | sonney2k: right | 21:12 |
| blackburn | but it should be actually I think :) | 21:12 |
| @sonney2k | while I don't mind adding thread safety for relevant cases we should do it only when really necessary | 21:12 |
| @sonney2k | blackburn, just think of someone calling train() of an SVM twice (in two threads...) | 21:13 |
| @sonney2k | lots of stuff will just kaboom | 21:13 |
| @sonney2k | gsomix, can you quickly measure say how long it takes to add say 100000 items via add() to CMap? | 21:15 |
| @sonney2k | gsomix, w/ and w/o locks? | 21:15 |
| blackburn | sonney2k: evaluations are open now :) | 21:20 |
| @sonney2k | blackburn, is this good or bad? | 21:22 |
| @sonney2k | blackburn, btw | 21:22 |
| blackburn | sonney2k: gsoc one I mean | 21:22 |
| @sonney2k | grep pthread shogun/ -r -l --include='*.cpp' | grep -v conv | wc -l | 21:22 |
| @sonney2k | 21 | 21:22 |
| @sonney2k | not so many ... | 21:22 |
| blackburn | I know | 21:22 |
| blackburn | sonney2k: and 4 done already | 21:23 |
| @sonney2k | yeah problem is that this is not a trivial task and it is easy to make mistakes | 21:25 |
| blackburn | sonney2k: it is not that bad actually | 21:30 |
| blackburn | I just paste helper thing into the place where pthread code was | 21:31 |
| alexlovesdata | @sonney2k: I will take a deep look into Viktors Code tomorrow | 21:32 |
| n4nd0 | sonney2k: around? | 21:35 |
| @sonney2k | blackburn, well look at shogun/features/DotFeatures.cpp | 21:36 |
| @sonney2k | actually that is not so difficult in this one | 21:36 |
| @sonney2k | blackburn, shogun/features/HashedWDFeaturesTransposed.cpp CHashedWDFeaturesTransposed::dense_dot_range_helper | 21:37 |
| @sonney2k | n4nd0, no :P | 21:37 |
| n4nd0 | sonney2k: I have planned some holidays for the next week, so I'll be some days off work; is there a big problem with it? | 21:38 |
| n4nd0 | sonney2k: I am certain it won't destroy the project's schedule | 21:39 |
| @sonney2k | n4nd0, np - just continue quickly when you are back | 21:40 |
| n4nd0 | sonney2k: yeah, I have also reserved more time for this week | 21:40 |
| n4nd0 | thank you | 21:40 |
| n4nd0 | is there any way we get to know if our mentor has already sent the midterm evaluation? | 21:41 |
| @sonney2k | n4nd0, no | 21:42 |
| -!- emrecelikten [~emre@213.153.218.163] has quit [Quit: Leaving.] | 21:43 | |
| gsomix | sonney2k, sorry. I'm back. | 21:44 |
| gsomix | >> quickly measure | is it actually needed? | 21:45 |
| gsomix | >> acquire_lock() / release_lock() | for lock/unlock? or for lock_init/lock_destroy? | 21:45 |
| @sonney2k | gsomix, what I had in mind is measuring CMap w/o locking | 21:47 |
| @sonney2k | (only add() function) | 21:47 |
| @sonney2k | and then w/ locking | 21:47 |
| gsomix | ok | 21:47 |
| @sonney2k | (from the patch you just sent) | 21:48 |
| @sonney2k | we can assume spinlocks so it should be pretty fast | 21:48 |
| @sonney2k | but lets see | 21:52 |
| blackburn | sonney2k: I am pretty sure locking of free lock is fast - no matter spinlock it is or mutex | 21:56 |
| @sonney2k | blackburn, my bet is that it is 10-100 times more expensive than x[i]=y | 21:58 |
| blackburn | sonney2k: my bet it is as fast :D | 21:59 |
| @sonney2k | gsomix, be the judge :D | 21:59 |
| -!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 248 seconds] | 21:59 | |
| @sonney2k | n4nd0, nico just sent me an email that he did the evals | 22:01 |
| blackburn | sonney2k: well why do you think so? | 22:01 |
| @sonney2k | gut feeling :D | 22:02 |
| @sonney2k | why do you think it is fast? | 22:02 |
| n4nd0 | sonney2k: yeah, thank you :) | 22:02 |
| blackburn | do you think any waits are here? | 22:02 |
| blackburn | it should be just atomic some bool=true | 22:02 |
| blackburn | well I don't know how to do that atomic though | 22:02 |
| -!- alexlovesdata [c25fae98@gateway/web/freenode/ip.194.95.174.152] has quit [Quit: Page closed] | 22:03 | |
| @sonney2k | n4nd0, btw before you leave just sent some early weekly report then - thanks | 22:09 |
| @sonney2k | blackburn, with spinlocks i think it really is just accessing some integer | 22:10 |
| @sonney2k | with the pthread stuff it is sth more complex | 22:10 |
| @sonney2k | but lets see | 22:10 |
| @sonney2k | gsomix, any news? | 22:10 |
| blackburn | with pthread stuff? | 22:10 |
| blackburn | isn't spinlock pthread stuff? | 22:10 |
| gsomix | sonney2k, measuring | 22:10 |
| @sonney2k | gsomix, btw I cannot access melange and fill out your evaluation... | 22:11 |
| gsomix | sonney2k, hmm, I'm not sure that I did it correctly. but... http://pastebin.com/WxCL76LK code and http://pastebin.com/eurxustX results | 22:22 |
| -!- ckwidmer [8ca3fe9d@gateway/web/freenode/ip.140.163.254.157] has joined #shogun | 22:24 | |
| blackburn | sonney2k: statistically insignificant | 22:24 |
| blackburn | :D | 22:24 |
| gsomix | oh, I had to make seven measurements... as said in the training manual for physics labs. :) | 22:26 |
| blackburn | lol | 22:26 |
| gsomix | I'm slowpoke... it seems that sonney2k gone to sleep | 22:28 |
| @sonney2k | gsomix, can you please measure only the add part with CTime and one more thing - are you using some on-demand cpu frequency scheduling? | 22:29 |
| @sonney2k | gsomix, turn that to a fixes freq in that case | 22:29 |
| @sonney2k | and for how to use CTime: CTime t; t.cur_time_diff(true); | 22:30 |
| @sonney2k | that's it | 22:30 |
| gsomix | >> turn that to a fixes freq in that case | I don't know how. :( | 22:30 |
| @sonney2k | gsomix, cpufreq-set -r -g performance | 22:36 |
| blackburn | heiko: ping | 22:43 |
| gsomix | sonney2k, blackburn http://pastebin.com/uyjZzj6P | 23:01 |
| gsomix | hm | 23:01 |
| blackburn | sonney2k: HAHA | 23:01 |
| gsomix | hm, weekly report | 23:03 |
| @sonney2k | gsomix, what is this print out? please use cur_time_diff() ... | 23:11 |
| blackburn | sonney2k: you looooost | 23:11 |
| @sonney2k | blackburn, c'mon the pthreaded stuff looks even faster than the other? | 23:11 |
| blackburn | sonney2k: INSIGNIFICANT :D | 23:12 |
| gsomix | sonney2k, okay. .____. | 23:12 |
| @sonney2k | gsomix, I think you used the clock_t based function... | 23:12 |
| gsomix | I used CTime with cur_runtime_diff | 23:13 |
| @sonney2k | blackburn, yes it seems to be fast enough though | 23:13 |
| gsomix | that "get time difference between start and NOW" | 23:13 |
| @sonney2k | gsomix, runtime_diff is not cur_time_diff | 23:13 |
| @sonney2k | I guess we need to modify the function names | 23:13 |
| @sonney2k | one is measuring cpu time | 23:14 |
| blackburn | yeah cur_time_diff | 23:14 |
| @sonney2k | the other wall clock time (in higher resolution) | 23:14 |
| blackburn | is *impossible* to remember | 23:14 |
| @sonney2k | blackburn, suggestions for a better name? | 23:14 |
| gsomix | oh, ok | 23:14 |
| @sonney2k | blackburn, gsomix it is like time <cmd> | 23:15 |
| @sonney2k | real0m0.003s | 23:15 |
| @sonney2k | user0m0.000s | 23:15 |
| @sonney2k | sys0m0.000s | 23:15 |
| gsomix | ok | 23:15 |
| @sonney2k | real == wall clock time | 23:15 |
| blackburn | sonney2k: I do not have any actually | 23:15 |
| @sonney2k | user == run_time in CTime | 23:15 |
| blackburn | my methods have bad names too :D | 23:15 |
| @sonney2k | I am no longer sure that the runtime based functions are ok... | 23:16 |
| @sonney2k | blackburn, well for a start - how about get_wall_clock_time() ? | 23:17 |
| @sonney2k | and get_consumed_cpu_time() ? | 23:17 |
| blackburn | better | 23:17 |
| @sonney2k | instead of | 23:17 |
| @sonney2k | get_curtime() | 23:17 |
| @sonney2k | and get_runtime() | 23:17 |
| @sonney2k | but I don't know how to name the functions measuring the difference between the saved start time and 'now' | 23:18 |
| blackburn | get_elapsed_clock_time | 23:19 |
| @sonney2k | get_elapsed_wall_clock_time() | 23:19 |
| @sonney2k | get_elapsed_cpu_time() | 23:19 |
| @sonney2k | I like that | 23:19 |
| heiko | blackburn what was it? | 23:22 |
| heiko | sorry was having diner :) | 23:22 |
| blackburn | heiko: no problem not actual I think | 23:22 |
| blackburn | I just didn't manage to find one xval function | 23:23 |
| blackburn | but now did | 23:23 |
| heiko | ok :) | 23:25 |
| CIA-18 | shogun: Heiko Strathmann master * r371587b / (9 files in 5 dirs): Merge pull request #634 from karlnapf/master (+8 more commits...) - http://git.io/XCtHgg | 23:25 |
| heiko | blackburn, I just added the covariance :) | 23:26 |
| heiko | will log off now, good night | 23:26 |
| blackburn | cool | 23:26 |
| -!- heiko [~heiko@host86-181-154-192.range86-181.btcentralplus.com] has left #shogun [] | 23:26 | |
| gsomix | blackburn, sonney2k http://pastebin.com/ri5Q6a5n | 23:30 |
| gsomix | .____. | 23:30 |
| @sonney2k | gsomix, what is the code? | 23:35 |
| blackburn | sonney2k: does pycon fits to present shogun? | 23:38 |
| gsomix | sonney2k, http://pastebin.com/fmGYTZfs | 23:38 |
| gsomix | tired | 23:43 |
| gsomix | need to sleep | 23:43 |
| gsomix | good night guys | 23:43 |
| -!- gsomix [~gsomix@88.200.242.178] has quit [Ping timeout: 244 seconds] | 23:49 | |
| --- Log closed Tue Jul 10 00:00:17 2012 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!