#pypy IRC log for Saturday, 2011-10-29

doloFor the links on http://doc.pypy.org/en/latest/extradoc.html (Papers), I find that the following two links are broken:00:02
doloTowards a Jitting VM for Prolog Execution, C.F. Bolz, M. Leuschel, D, Schneider (http://www.stups.uni-duesseldorf.de/publications/bolz-prolog-jit.pdf)00:03
tlynn_ (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) left irc: Ping timeout: 240 seconds00:03
doloAutomatic JIT Compiler Generation with Runtime Partial Evaluation (Master Thesis), C.F. Bolz (http://www.stups.uni-duesseldorf.de/thesis/final-master.pdf)00:03
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/92200:10
etrepum (~bob@accessnat4.mochimedia.net) left irc: Quit: etrepum00:11
etrepum (~bob@accessnat4.mochimedia.net) joined #pypy.00:12
etrepum (~bob@accessnat4.mochimedia.net) left irc: Client Quit00:13
etrepum (~bob@accessnat4.mochimedia.net) joined #pypy.00:13
Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 244 seconds00:15
rekamso (~textual@216.190.29.42) left irc: Quit: .00:15
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Ping timeout: 255 seconds00:21
etrepum (~bob@accessnat4.mochimedia.net) left irc: Quit: etrepum00:29
etrepum (~bob@accessnat4.mochimedia.net) joined #pypy.00:29
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/181200:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/54600:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/59400:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/107400:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/78000:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/141200:49
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/67800:49
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/24400:49
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.00:50
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer00:52
tilgovi (~randall@112.sub-75-208-51.myvzw.com) joined #pypy.00:54
tilgovi (~randall@112.sub-75-208-51.myvzw.com) left irc: Changing host00:54
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.00:54
zooko (~user@97-118-92-53.hlrn.qwest.net) joined #pypy.01:05
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Remote host closed the connection01:06
lac_ (~quassel@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy.01:06
lac (~quassel@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 244 seconds01:07
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep.01:16
mcdonc (~mcdonc@cabana.palladion.com) left irc: Ping timeout: 258 seconds01:17
timonatorare the new results on speed.pypy.org still due to a bug or the system being used by some other process? they look pretty terrible01:18
etrepum (~bob@accessnat4.mochimedia.net) left irc: Ping timeout: 248 seconds01:25
mcdonc (~mcdonc@cabana.palladion.com) joined #pypy.01:30
berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) joined #pypy.01:30
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/11901:46
Moku (~John@osbk-4db064c6.pool.mediaWays.net) joined #pypy.01:52
Shinita (~John@osbk-4d087ff6.pool.mediaWays.net) left irc: Ping timeout: 240 seconds01:53
Nick change: Moku -> Guest143701:53
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/24401:57
Arach (~arach@2.95.26.12) left irc: Ping timeout: 276 seconds02:00
canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving.02:11
Arach (~arach@2.95.28.85) joined #pypy.02:12
DasIch (~dasich@p4FFDEBE3.dip.t-dialin.net) left irc: Ping timeout: 260 seconds02:12
verte (~verte@python/site-packages/verte) joined #pypy.02:15
berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) left irc: Remote host closed the connection02:18
DasIch (~dasich@p4FFDEFAD.dip.t-dialin.net) joined #pypy.02:24
zooko (~user@97-118-92-53.hlrn.qwest.net) left irc: Ping timeout: 252 seconds02:24
twomashi (~Adium@201-213-122-142.net.prima.net.ar) joined #pypy.02:27
zooko (~user@97-118-92-53.hlrn.qwest.net) joined #pypy.02:32
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy02:34
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy.02:37
twomashianyone around?02:45
vertebetter than being square02:46
twomashiba ding02:46
vertewhat's up?02:47
twomashiim trying to use jitviewer to see how to speed up a module02:47
twomashibut i dont. understand. anything02:48
twomashii want to speed up the json decoding02:48
verteoh, cool02:48
vertewell, there was a blog post on using the jitviewer, iirc02:49
twomashithe recent one, on the json encoder?02:49
twomashii have it working, and im looking at the output of my test02:49
twomashiits interpreting it that im stuck on.02:49
vertehttp://morepypy.blogspot.com/2011/08/visualization-of-jitted-code.html02:50
twomashiah cool..02:52
twomashiwell, unfortunately the jit viewer isnt giving me the name of the two most frequently run items02:53
verteare they fairly short? they might be top-level or internal to the interpreter02:54
twomashiys02:54
twomashithe top one is 16 lines of assembly with 4 guards02:54
Alex_Gaynora regular expression perhaps?02:55
vertedoes it use regex? I didn't think it did02:56
twomashiit does02:56
zooko (~user@97-118-92-53.hlrn.qwest.net) left irc: Remote host closed the connection02:56
zooko (~user@97-118-92-53.hlrn.qwest.net) joined #pypy.02:56
twomashialso the viewer is throwing an exception02:57
vertepretty wierd, considering parsing json only requires lookahead for character escapes02:57
twomashiit uses it to distinguish between int and float02:58
zooko (user@97-118-92-53.hlrn.qwest.net) left #pypy.02:58
twomashiwhich is fairly sane in some cases02:58
Alex_Gaynortwomashi: you still can based on the variable names, but we switched to symbols to make it more user friendly02:58
verteah02:58
Action: Alex_Gaynor is moderately curious who can't understand int_mul and float_mul; but taht's life02:59
gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Remote host closed the connection03:02
twomashiis it normal to see so many calls to setarrayitem_gc?03:04
twomashitheres dozens of them after a return statement03:04
Arach (~arach@2.95.28.85) left irc: Ping timeout: 260 seconds03:04
vertevery common03:04
Alex_Gaynorthat's what happens when you compile a single function, rather than a loop03:04
Alex_Gaynornothing you can do about it03:04
vertethey may be clearing the valuestack03:05
Alex_Gaynoryou could del every local variable possibly, feels evil though03:06
verteoh, the locals too. good point.03:07
Alex_Gaynorthe valuestack will be empty, as long as teh final return value is just a variable, not a complex expression03:07
twomashiit does, but on the other hand, this is microoptimisation...03:08
Alex_Gaynoryeah, I'd look around elsewhere first :/03:08
timonatorhum, how would return a_complex_expression differ from foo = a_complex_expression; return foo; ?03:10
timonatoroh, well, the python opcodes would be slightly different i suppose?03:10
Alex_Gaynoryou could del any local variables between them03:10
Alex_Gaynorotherwise, not at all03:11
verteat the opcode level there's not much optimisation that goes on03:11
jonvk (~jonathan@bas1-montreal29-1279650546.dsl.bell.ca) left irc: Remote host closed the connection03:12
timonatormhm, so what you're saying is, that del-ing every local variable before the return yields close to no performance improvement at all, yes?03:14
Alex_GaynorI have noi dea what improvement it yields03:14
Alex_Gaynorsomeone should measure it03:14
timonatorwell, it's been nice chatting with y'all03:14
Action: timonator runs away03:14
twomashiif s[end:end + 1] != ':':03:16
twomashiSLICE_303:16
tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) joined #pypy.03:16
tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) left irc: Changing host03:16
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.03:16
twomashithat might be a good target right?03:16
Arach (~arach@78-106-251-8.broadband.corbina.ru) joined #pypy.03:17
Alex_Gaynoryeah, that'll prbably generate a call, I'd rewrite it as an explicit length check + indexed lookup03:17
Alex_Gaynorif len(s) > end and s[end] !+ ':'03:17
timonatoris this something pypy could reasonably do by itself? (and is it worthwhile?)03:19
Alex_Gaynoryes, probably03:20
verteit'd be worthwhile to do in the jit, where it isn't so sensible to do slice string strategies in general03:21
Alex_Gaynoractually I think it does03:22
Alex_Gaynorthe JIT seems to automatically do it for length 1 ones03:22
twomashimight be easy to detect if the slice is only used for comparative purposes (?)03:22
twomashinevermind :p03:24
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Computer has gone to sleep.03:38
htoothrot (~mux@unaffiliated/htoothrot) left irc: Ping timeout: 260 seconds03:38
etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy.03:39
mcdonc_ (~mcdonc@cabana.palladion.com) joined #pypy.03:41
mcdonc (~mcdonc@cabana.palladion.com) left irc: Ping timeout: 258 seconds03:42
price (~price@LINERVA.MIT.EDU) joined #pypy.03:48
mcdonc_ (~mcdonc@cabana.palladion.com) left irc: Remote host closed the connection03:49
mcdonc_ (~mcdonc@cabana.palladion.com) joined #pypy.03:49
verte (~verte@python/site-packages/verte) left irc: Quit: ~~~ Crash in JIT!04:04
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/181204:09
Arach (~arach@78-106-251-8.broadband.corbina.ru) left irc: Ping timeout: 245 seconds04:34
Arach (~arach@FTTB-dynamic-79.104.203.92.ranetka.ru) joined #pypy.04:47
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/141205:10
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/59405:19
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy05:30
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/107405:30
gutworth (~benjamin@dhcp-128-36-89-186.central.yale.edu) left irc: Ping timeout: 255 seconds05:30
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/78005:33
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/54605:35
buggaboo (~jasm@535316B2.cm-6-4a.dynamic.ziggo.nl) joined #pypy.05:42
mat^2 (~mathias@212.130.113.35) joined #pypy.05:53
asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer05:59
kenaan12alex_gaynor jit-raw-array-of-struct 11bb85820f08a3 15/pypy/jit/: random changes06:03
kenaan12alex_gaynor default 11753627dbee28 15/pypy/module/pypyjit/test_pypy_c/test_containers.py: fix this test for 32-bit machines06:03
twomashi (~Adium@201-213-122-142.net.prima.net.ar) left irc: Quit: Leaving.06:17
lac_ (~quassel@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 244 seconds06:48
Nick change: nopf_ -> nopf07:21
buggaboo (~jasm@535316B2.cm-6-4a.dynamic.ziggo.nl) left irc: Remote host closed the connection07:52
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/67807:56
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 260 seconds08:15
mat^2 (~mathias@212.130.113.35) joined #pypy.08:20
fijalhi08:27
tlynn (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) joined #pypy.08:29
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 260 seconds08:35
mat^2 (~mathias@212.130.113.35) joined #pypy.08:36
mat^2 (~mathias@212.130.113.35) left irc: 09:02
Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy.09:09
mat^2 (~mathias@212.130.113.35) joined #pypy.09:22
Hodgestar (~Hodgestar@gluon.za.net) left irc: Ping timeout: 260 seconds09:44
Hodgestar (~Hodgestar@gluon.za.net) joined #pypy.09:45
JaRoel (~jaroel|4d@sink.jaroel.nl) joined #pypy.09:51
markpearse (d988665a@gateway/web/freenode/ip.217.136.102.90) joined #pypy.10:02
JaRoel (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 260 seconds10:30
JaRoel|4d (~jaroel|4d@2001:980:7133:1:48ea:28c5:907d:132c) joined #pypy.10:34
stakkars_ (~tismer@p5DDB5E98.dip.t-dialin.net) joined #pypy.10:37
stakkars (~tismer@p5DDB5E98.dip.t-dialin.net) left irc: Read error: Connection reset by peer10:38
Nick change: stakkars_ -> stakkars10:38
kkris (~kris@80-123-43-84.adsl.highway.telekom.at) joined #pypy.10:58
k_bx (~k_bx@94.244.19.62) joined #pypy.11:02
fschwidom (~fschwidom@109.250.152.206) joined #pypy.11:12
htoothrot (~mux@unaffiliated/htoothrot) joined #pypy.11:18
k_bx (~k_bx@94.244.19.62) left irc: Remote host closed the connection11:35
k_bx (~k_bx@94.244.19.62) joined #pypy.11:35
k_bx (~k_bx@94.244.19.62) left irc: Remote host closed the connection11:36
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Ping timeout: 252 seconds11:41
derdon (~derdon@pD9E1CF60.dip.t-dialin.net) joined #pypy.11:44
EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) left irc: Read error: Connection reset by peer11:46
mcdonc_ (~mcdonc@cabana.palladion.com) left irc: Ping timeout: 258 seconds11:51
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) joined #pypy.12:14
PiotrSikora (~none@nginx/adept/piotrsikora) left irc: Excess Flood12:15
ojii (~ojii@75-135.79-83.cust.bluewin.ch) joined #pypy.12:17
mcdonc_ (~mcdonc@cabana.palladion.com) joined #pypy.12:18
PiotrSikora (~none@nginx/adept/piotrsikora) joined #pypy.12:18
Sho_ (~EHS1@kde/hein) left irc: Quit: Stop leaking memory like it's going out of fashion.12:28
JaRoel|4d (~jaroel|4d@2001:980:7133:1:48ea:28c5:907d:132c) left irc: Ping timeout: 244 seconds12:38
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) joined #pypy.12:39
JaRoel (~jaroel|4d@sink.jaroel.nl) joined #pypy.12:43
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 240 seconds12:44
JaRoel (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 260 seconds12:50
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) joined #pypy.12:51
JaRoel (~jaroel|4d@sink.jaroel.nl) joined #pypy.12:54
legogris (~legogris@c80-217-200-244.bredband.comhem.se) joined #pypy.12:54
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep.12:56
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 258 seconds12:57
k_bx (~k_bx@94.244.19.62) joined #pypy.12:58
k_bxMorning everyone! I was playing with this article about dropbox python web site optimizations http://tech.dropbox.com/?p=89 and what I saw strange is that in this simple code http://paste.pocoo.org/show/499952/ when you put WHITELIST as a set() it actually works slower than if you just put it as a string (without set()). Without set time python main.py gives me 4.855s, while set() variant gives 5.540s12:59
gutworth (~benjamin@128.36.89.186) joined #pypy.12:59
k_bx(on pypy, of course)13:00
k_bxWhich is strange, I guess, since set()'s "in" check should be faster (as I thought)13:00
mcdonc_ (~mcdonc@cabana.palladion.com) left irc: Ping timeout: 258 seconds13:01
gutworthis this also true on cpython?13:03
k_bxhere's a smaller example http://paste.pocoo.org/show/499955/13:05
k_bxno, on cpython set() version is significantly faster13:05
k_bxcpython: http://paste.pocoo.org/show/499956/13:06
JaRoel (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 256 seconds13:06
gutworththis is mostly because sets are not as nicely optimized in the jit as strings13:07
JaRoel (~jaroel|4d@sink.jaroel.nl) joined #pypy.13:08
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) joined #pypy.13:09
JaRoel (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 252 seconds13:13
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.13:21
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Client Quit13:21
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.13:22
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Client Quit13:22
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.13:34
k_bx (~k_bx@94.244.19.62) left irc: Remote host closed the connection13:44
fschwidom (~fschwidom@109.250.152.206) left irc: Remote host closed the connection13:56
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/13:57
berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) joined #pypy.14:24
berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) left irc: Remote host closed the connection14:24
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 276 seconds14:31
hobu (~hobu@prometheus.hobu.net) left irc: Changing host14:33
hobu (~hobu@osgeo/member/hobu) joined #pypy.14:33
hobu (~hobu@osgeo/member/hobu) left irc: Excess Flood14:33
hobu (~hobu@osgeo/member/hobu) joined #pypy.14:33
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) joined #pypy.14:35
shiyas (~shiyasras@cpc2-aztw18-0-0-cust104.aztw.cable.virginmedia.com) joined #pypy.14:39
JaredW (~JaredW@118.142.17.238) left irc: Read error: Operation timed out14:41
JaredW (~JaredW@118.142.17.238) joined #pypy.14:41
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep.14:48
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) joined #pypy.14:49
Arach (~arach@FTTB-dynamic-79.104.203.92.ranetka.ru) left irc: Ping timeout: 260 seconds15:01
Unhelpful (~quassel@rockbox/developer/Unhelpful) left irc: Ping timeout: 240 seconds15:08
Arach (~arach@2.95.176.212) joined #pypy.15:17
cpets (4f75d76f@gateway/web/freenode/ip.79.117.215.111) joined #pypy.15:26
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep.15:26
mcdonc_ (~mcdonc@cabana.palladion.com) joined #pypy.15:27
mat^2 (~mathias@212.130.113.35) joined #pypy.15:31
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) joined #pypy.15:50
zooko (~user@97-118-92-53.hlrn.qwest.net) joined #pypy.15:51
aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep.16:02
mcdonc_ (~mcdonc@cabana.palladion.com) left irc: Remote host closed the connection16:09
hobu (~hobu@osgeo/member/hobu) left irc: Excess Flood16:10
hobu (~hobu@osgeo/member/hobu) joined #pypy.16:10
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy.16:14
Unhelpful (~quassel@rockbox/developer/Unhelpful) joined #pypy.16:23
Arfrever (~Arfrever@apache/committer/Arfrever) joined #pypy.16:27
mcdonc_ (~mcdonc@cabana.palladion.com) joined #pypy.16:35
shiyas (~shiyasras@cpc2-aztw18-0-0-cust104.aztw.cable.virginmedia.com) left irc: Quit: shiyas16:39
poincare101 (~Dhaivat_P@CPE-65-27-119-146.new.res.rr.com) joined #pypy.16:52
poincare101Hello everyone. I currently have flask running under python 2.7, how can I start using pypy instead? Where how are modules supposed to be saved?16:52
Unhelpful (~quassel@rockbox/developer/Unhelpful) left irc: Ping timeout: 240 seconds16:54
ojiithe pypy/benchmarks are not mirrored to github, are they?16:56
rguillebertpoincare101, pypy is (theoretically) a direct replacement to cpython17:11
qbproger (~qbproger@c-76-104-159-248.hsd1.wa.comcast.net) joined #pypy.17:12
rguillebertso you just need to run your code directly17:12
rguillebertexcept if you have C extensions17:12
fijalojii: nope17:12
fijalpoincare101: install it into a virtualenv17:13
ojiifijal, meh cloned the hg now17:13
ojiifijal, can you tell me how i can run the benchmarks locally?17:13
qbprogerfijal: I saw you were doing json perf, did you look at https://github.com/esnme/ultrajson?17:13
ojiiwanna see if i can write a benchmark of my own :D17:14
Alex_Gaynorfijal: what's up with all the crashes on pypy-c-ojit-no-jit17:17
derdon (~derdon@pD9E1CF60.dip.t-dialin.net) left irc: Remote host closed the connection17:28
JaRoel (~jaroel|4d@2001:980:7133:1:39b2:f5cc:8a54:7bbc) joined #pypy.17:32
fijalojii: ./runner.py --help17:32
ojiifijal, yep figured it out, thanks17:33
fijalqbproger: not really, why?17:33
fijalAlex_Gaynor: no idea?17:33
Alex_Gaynorfijal: I guess someone should look :/17:33
qbprogerfijal: i was looking at their benchmarks... looked pretty impressive17:33
fijalqbproger: you mean numbers or benchmarks themselves?17:33
kenaan12alex_gaynor default 11a257711072e5 15/pypy/jit/metainterp/optimizeopt/virtualize.py: fix so these test don't rely on dict ordering17:34
qbprogernumbers17:34
kenaan12alex_gaynor default 11a3a5ac0a2daf 15/pypy/module/pypyjit/test_pypy_c/test_containers.py: merged default17:34
asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.17:34
qbprogerfijal: though the actual benchmarks may be interesting too17:34
fijalAlex_Gaynor: does not look very good17:35
Alex_Gaynorfijal: it's probably a pretty shallow bug, right?17:35
fijalno idea17:35
fijalwhat did we change anyway?17:36
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) left irc: Ping timeout: 260 seconds17:36
Alex_Gaynordunno, that was the day you merged the finalizer changes right?17:36
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy17:36
Alex_Gaynorwhat GC does ojit nojit use?17:36
fijalminimark I guess17:36
Alex_Gaynorfijal: it uses the hybrid GC, I guess the bug is probably there17:36
fijalmodule tests are also broken17:37
fijalhttp://buildbot.pypy.org/summary17:37
Alex_Gaynorfijal: that's wiht -A, so no surprise17:37
Alex_GaynorI claim it's probably related to hybrid GC17:38
fijalyop17:38
fijalI might have broken it for all I know17:38
Alex_GaynorI consider it likely :)17:39
Alex_Gaynorfijal: also, jvm/clr tests are broken for strip :)17:40
ojiifijal, tried to write a benchmark for that dulwich slowness i was talking about a few days ago, but for some reason it doesn't really show in the benchmark i wrote. http://dpaste.com/643602/17:42
ojiialso since i had the brilliant idea to use pypy.git as the repo to benchmark it is very slow :D17:44
durin42 (~durin@adium/durin42) left irc: Remote host closed the connection17:50
zooko (user@97-118-92-53.hlrn.qwest.net) left #pypy ("#tahoe-lafs").17:53
htoothrot (~mux@unaffiliated/htoothrot) left irc: Ping timeout: 260 seconds18:00
shiyas (~shiyasras@cpc2-aztw18-0-0-cust104.aztw.cable.virginmedia.com) joined #pypy.18:02
ronnyojii: afair the combination of string-slicing, regex scans, dicts and other stuff is not quite as effective as a c solution18:03
ronnyit could be a lot better if struct could deal with zero terminated strings18:04
kkris (~kris@80-123-43-84.adsl.highway.telekom.at) left irc: Ping timeout: 252 seconds18:05
ojiironny, no c here, just python vs pypy18:09
ojiiremoved the c modules in dulwich in both cases18:09
ojiistill in my real application python is 2x faster18:09
ojii1.918s vs 0.926s18:10
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.18:10
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Client Quit18:11
gutworthAlex_Gaynor: is it snowing there?18:11
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.18:12
Alex_Gaynorgutworth: it did the other day, let me glance out the window18:13
Alex_Gaynorgutworth: nope ATM our skies are clear18:13
Alex_Gaynorojii: copy paste your application into the benchmark18:13
ojiiAlex_Gaynor, ugh more dependency hacking :(18:14
Alex_Gaynorojii: FWIW have you tried running your app multiple times in the same process, just to make sure it isn't a warmup time issue?18:14
ojiiAlex_Gaynor, https://github.com/ojii/gitstats.ep.io/blob/master/src/run.py18:14
stakkars_ (~tismer@dslb-088-072-246-130.pools.arcor-ip.net) joined #pypy.18:15
Alex_Gaynorgutworth: snowing down there?18:15
ojiinot sure how i'd back all that including the 3 dependencies + 2 modules i wrote into a benchmark18:16
Alex_Gaynorojii: have you established that the dulwich part of your application is the slow part?18:16
ojiino18:16
ojiialso it's not slow, i was just playing around to see how it performs under pypy18:16
ojiiand was surprised to see that initially it was 4x slower18:16
Alex_Gaynorslower.  it's the same as slow to me :)18:17
ojiiha18:17
Alex_GaynorI'm going to run out and get lunch now, be back shortly18:17
ojiihrm i should eat diner18:17
ojii*dinner18:17
ronnyojii: interesting18:17
ronnyojii: that basically means you have no-jit speed18:18
ronan__ (~ronan@host81-132-98-104.range81-132.btcentralplus.com) joined #pypy.18:18
htoothrot (~mux@unaffiliated/htoothrot) joined #pypy.18:18
gutworthAlex_Gaynor: yes18:19
ronan (~ronan@host86-151-196-170.range86-151.btcentralplus.com) left irc: Ping timeout: 248 seconds18:19
gutworthI guess we're not close enough to share weather18:19
kkris (~kris@80-123-43-227.adsl.highway.telekom.at) joined #pypy.18:20
EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) joined #pypy.18:24
stakkars_ (~tismer@dslb-088-072-246-130.pools.arcor-ip.net) left irc: Quit: schnarch18:24
ojiii think i found the slow part: http://dpaste.com/643622/18:27
Alex_Gaynorojii: there we go!18:29
Alex_Gaynorojii: my only concern is that the times might be so short it's just noise, can you create a longer running one?18:29
ojiiAlex_Gaynor, how?18:29
Alex_Gaynorstick it in a loop?18:30
ojiiit is18:30
ojii1 sec18:30
Alex_Gaynorcan you paste the benchmark?18:30
ojiihttp://paste.pocoo.org/show/500111/18:30
ojiiha forgot to change a comment string there :D18:31
Alex_Gaynorojii: also I think this always produces no results doesn't it?18:33
ojiihm?18:33
Alex_Gaynoryou switched the start and end18:33
ojiierm fail...18:33
ojiiAlex_Gaynor, http://paste.pocoo.org/show/500114/18:34
Alex_Gaynorojii: can you run at -n25000 just for reference?18:35
ojiino such option: -n18:35
Alex_Gaynorojii: uh sorry, can you change wherever you set that it does 5k iterations to 25k?18:36
ojiidid it with the iteration_scaling thing18:36
Alex_Gaynoryeah can you bump up the iteraitosn, just for comparison?18:36
ojiiyep18:36
ojiirunning...18:36
ojiiAlex_Gaynor, there you go: http://paste.pocoo.org/show/500117/18:38
Alex_Gaynorok so not just warmup time18:38
ojiibtw according to that tool, pypy has 153 total contributors :D18:42
Action: ojii loves statistics18:43
nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.18:43
Alex_Gaynorojii: http://paste.pocoo.org/show/500119/18:44
Alex_GaynorI'm using trunk FWIW18:47
exarkunojii: but what's the slope of 3 month running mean of contributors?18:56
ojiihrm?18:57
Action: exarkun loves statistics too18:57
ojiiexarkun, i'll hopefully finish the bit of that project that makes it display more than one repo18:58
ojiiexarkun, if there's more different stats you'd want, fork it :D18:59
Nick change: Gulaway -> Gulopine19:26
Trundle (~andy@82.113.99.135) joined #pypy.19:36
Trundle (~andy@82.113.99.135) left irc: Changing host19:36
Trundle (~andy@python/site-packages/trundle) joined #pypy.19:36
stakkars_ (~tismer@dslb-088-072-246-130.pools.arcor-ip.net) joined #pypy.19:37
asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Read error: Connection reset by peer19:39
asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.19:43
amaury (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy.19:45
cpets (4f75d76f@gateway/web/freenode/ip.79.117.215.111) left irc: Ping timeout: 265 seconds19:52
fijalojii: I'll look tomorrow19:57
fijalmaybe :)19:57
ojiifijal, don't worry19:57
ojiias i said, not really a big deal19:57
fijalthis is mostly _structseq being applevel and not really nice for the JIT19:57
ojiiep.io doesn't run pypy yet anyway, so....19:57
fijalthey do, but not officially so ;-)19:57
ojiiyea that's what i meant20:00
fijalanyway20:01
fijalstructseq is my favorite suspect for now20:02
Alex_Gaynorfijal: we should tell carl he can write a paper by letting us have app level look_inside_iff20:06
twomashi (~Adium@201-213-122-142.net.prima.net.ar) joined #pypy.20:07
amaury (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Read error: Operation timed out20:09
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 276 seconds20:12
fijalAlex_Gaynor: mess20:13
Alex_Gaynorfijal: definitely :)20:13
shiyas (~shiyasras@cpc2-aztw18-0-0-cust104.aztw.cable.virginmedia.com) left irc: Quit: shiyas20:16
asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer_20:28
cpets (4f75d76f@gateway/web/freenode/ip.79.117.215.111) joined #pypy.20:30
Alex_Gaynorgutworth: snowing :)20:31
RhyoliteAlex_Gaynor: it has been snowing here for awhile.  Snow in October.  Ugh!20:32
Alex_GaynorRhyolite: we had first snow a few days ago, it's pouring down again now though20:32
Rhyolitea colleague drove back from Poughkeepsie on Thursday evening and said it was a mess20:33
twomashiHi all20:34
asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.20:35
amaury (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy.20:36
fijal (~fijal@78.9.163.32) left irc: Ping timeout: 245 seconds20:37
Alex_GaynorRhyolite: I can imagine20:40
twomashihi alex20:41
twomashido you have time to take a look at a jitviewer output for json decode and see if you can identify any bottlenecks?20:41
Alex_Gaynornot especially, have you looked at profiles from cProfile?20:42
twomashii ran it using cProfile but it ran for so long that I assumed the module didn't work properly using the pypy jit20:43
Alex_Gaynorit should20:44
Alex_Gaynorit is slower though20:44
twomashiah, perhaps i was using reguar profile20:44
twomashiso the cProfile output applies to jitted code?20:45
Alex_Gaynoryes20:46
twomashiok great20:46
twomashithanks20:46
durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) joined #pypy.20:49
durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) left irc: Changing host20:49
durin42 (~durin@adium/durin42) joined #pypy.20:49
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Computer has gone to sleep.21:05
poincare101 (~Dhaivat_P@CPE-65-27-119-146.new.res.rr.com) left irc: Read error: Connection reset by peer21:07
poincare101 (~Dhaivat_P@CPE-65-27-119-146.new.res.rr.com) joined #pypy.21:07
Ademan (~dan@adsl-71-141-242-47.dsl.snfc21.pacbell.net) joined #pypy.21:26
asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Ping timeout: 255 seconds21:27
asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.21:29
tlynn (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) left irc: Ping timeout: 260 seconds21:39
ojiiexarkun, https://gitstats.ep.io/pypy/21:40
Alex_Gaynorojii: now is there a way to compare that to other projects?21:41
ojii(for new authors bars you might have to zoom)21:41
ojiiAlex_Gaynor, fork it!21:41
ojii:D21:41
Alex_Gaynorojii: yeah, with my enormous free time ;)21:41
ojiiexactly!21:41
ojiibut yea, that's an idea i had in mind with this21:42
ojiifirst i gotta find a nicer way to auto-generate these so i don't have to run a manual command each time i want to update a graph21:42
ojiiand add more graphs :D21:42
Alex_Gaynorlast time I did the math pypy's had an average of 16 commits per day for the last 8 years or something.  pretty sure e.g. django isn't even close21:42
ojiiAlex_Gaynor, should I run django on it?21:43
Alex_Gaynorsure, why not21:43
ojiibecause it'll take forever :D21:43
ojiipypy took looong21:43
Alex_Gaynordjango has muuuuch less commits than pypy21:43
Alex_Gaynorlike half as much21:43
ojiiand building...21:43
danchrFWIW 'hg churn' gives you some nice statistics too  and it's reasonably fast with the -c option21:43
ojiidulwich is sweet21:44
ojiicompletely undocumented but still fun21:44
ojiiany other git repos anyone wants? (no, i won't do linux)21:44
Alex_Gaynorlinux?21:44
Alex_Gaynorgit21:44
ojiithose would take FOREVER to build21:44
Alex_Gaynortry running THOSE on pypy and see if it's faster :D21:45
ojiiand probably crash the browser with the amount of js they'd generate21:45
Alex_Gaynorpansy browsers!21:45
ojiihttps://gitstats.ep.io/django/21:45
danchrojii: are you running the against Git repositories?21:45
ojiidanchr, that's git repository statistics, yes21:45
Alex_Gaynorlove the october 2010 increase21:45
Alex_Gaynor:)21:45
ojiiof authors?21:46
danchryou might want to try running it against Mercurial repos  DAG walking is supposedly quite slow in Git21:46
stakkars_ (~tismer@dslb-088-072-246-130.pools.arcor-ip.net) left irc: Quit: schnarch21:46
Alex_Gaynorojii: yeah, that was when myself and like 6 other people got commit :)21:46
ojiiyep21:46
Trundle (~andy@python/site-packages/trundle) left irc: Remote host closed the connection21:46
ojiialso like the spike in authors pypy got in nov 201021:46
ojiinot spike, but increased speed in 'cumulative authors'21:47
ronan__ (~ronan@host81-132-98-104.range81-132.btcentralplus.com) left irc: Ping timeout: 240 seconds21:47
Action: ojii looks for more interesting git repos21:48
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/12021:49
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/92321:49
bbot2Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/69621:49
amaury (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 240 seconds21:51
twomashi (~Adium@201-213-122-142.net.prima.net.ar) left irc: Quit: Leaving.22:08
cpets (4f75d76f@gateway/web/freenode/ip.79.117.215.111) left irc: Ping timeout: 265 seconds22:09
twomashi (~Adium@201-213-122-142.net.prima.net.ar) joined #pypy.22:09
tos9 (~tos9@unaffiliated/tos9) left irc: Ping timeout: 260 seconds22:10
mcdonc_ (~mcdonc@cabana.palladion.com) left irc: Remote host closed the connection22:22
tos9 (~tos9@unaffiliated/tos9) joined #pypy.22:22
kenaan12matthp numpy NDimArray 11a2f79669b123 15/pypy/module/micronumpy/interp_numarray.py: NDimSlice22:35
kenaan12mattip numpy NDimArray 114ef01c9532c9 15/pypy/module/micronumpy/interp_numarray.py: print of ndarray works (kindof)22:35
kenaan12mattip numpy NDimArray 11827d5d5d5218 15/pypy/module/micronumpy/: Slices work for get and set22:35
kenaan12mattip numpy NDimArray 1129e96633349c 15/: Merge22:35
Action: pjenvey didn't even know branch names could have whitespace22:35
Matti_ (~chatzilla@bzq-109-67-120-155.red.bezeqint.net) joined #pypy.22:39
Matti_Hello.22:39
Alex_GaynorMatti_: FYI fijal started working on multi-dim as well in a seperate branch22:39
Matti_I have just comitted some updates to numpy NDim, now slices work.22:40
Matti_Ahh, can we merge our stuff somehow? I will have to check out what fijal did.22:40
Matti_For what it's worth, in my branch you can do things like22:41
Matti_import numpy22:41
Matti_a=numyp.zeros((3,4))22:41
Alex_Gaynorhis branch is `numpy-multidim`, I'd send a mail to pypy-dev and coordinate with him22:41
Alex_Gaynorobviously we don't want any duplicate work22:41
Alex_GaynorI think he's probably in bed ATM22:41
Matti_a[:,0:2]=522:41
Alex_GaynorI think that works on his branch as well, but I'm only kinda following, please send a mail to the mailing list22:42
Alex_Gaynor:)22:42
Matti_OK, gnight22:43
Matti_ (chatzilla@bzq-109-67-120-155.red.bezeqint.net) left #pypy.22:43
timonatori wonder when bool-arrays for slicing and comparison arrays will work22:49
jonanin (~jonanin@24-183-50-140.dhcp.mdsn.wi.charter.com) left irc: Remote host closed the connection22:52
JaRoel (~jaroel|4d@2001:980:7133:1:39b2:f5cc:8a54:7bbc) left irc: Ping timeout: 244 seconds23:02
JaRoel|4d (~jaroel|4d@2001:980:7133:1:d460:1c39:8d38:b04d) joined #pypy.23:05
etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Read error: Operation timed out23:12
tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) joined #pypy.23:15
tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) left irc: Changing host23:15
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.23:15
kkris (~kris@80-123-43-227.adsl.highway.telekom.at) left irc: Remote host closed the connection23:19
bbot24Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/69623:35
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1813 [12pcl, 1.6]23:45
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/679 [12pcl, 1.6]23:45
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/535 [12pcl, 1.6]23:45
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1075 [12pcl, 1.6]23:45
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/547 [12pcl, 1.6]23:45
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/245 [12pcl, 1.6]23:45
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/245 [12pcl, 1.6]23:45
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/535 [12pcl, 1.6]23:46
Arfrever (~Arfrever@apache/committer/Arfrever) left irc: Quit: Ex+re23:48
etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy.23:55
--- Sun Oct 30 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!