#pypy on 2017-10-24 — irc logs at freenode.irclog.whitequark.org

2017-10-17 08:34 antocuni changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | "PyPy: the Gradual Reduction of Magic (tm)"

00:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/708 [default]

00:05 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/3367 [default]

00:07 <bbot2> Success: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4286 [default]

00:23 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

00:30 tbodt has joined #pypy

00:32 marr has quit [Ping timeout: 264 seconds]

00:32 tbodt has quit [Client Quit]

00:46 ceridwen has quit [Ping timeout: 252 seconds]

00:50 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-linux-armhf-raspbian/builds/1575

00:50 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-app-level-linux-armhf-raspbian/builds/1503

00:50 ceridwen has joined #pypy

00:50 tbodt has joined #pypy

00:50 lritter_ has joined #pypy

00:53 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-app-level-linux-armhf-raspbian/builds/1503

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4287 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/3368 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5017 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/6317 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/5525 [py3.5]

01:07 __Yiota has joined #pypy

01:12 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armhf-raring/builds/1699

01:13 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

01:15 tbodt has joined #pypy

01:20 tbodt has quit [Ping timeout: 258 seconds]

01:23 jcea has quit [Quit: jcea]

01:25 __Yiota has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

01:47 __Yiota has joined #pypy

01:57 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-linux-armel/builds/1835

02:06 kipras is now known as kipras`away

02:08 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5017 [py3.5]

02:26 <bbot2> Success: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/708 [default]

02:30 hvariant has joined #pypy

02:31 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4287 [py3.5]

02:40 lritter__ has joined #pypy

02:41 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/6317 [py3.5]

02:44 lritter_ has quit [Ping timeout: 240 seconds]

02:55 ArneBab has joined #pypy

02:58 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armhf-raspbian/builds/1704

02:58 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-armhf-raspbian/builds/1450

02:59 ArneBab_ has quit [Ping timeout: 240 seconds]

03:00 <bbot2> Started: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/2370

03:00 <bbot2> Started: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/3162

03:02 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/5525 [py3.5]

03:08 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-s390x/builds/743 [default]

04:53 forgottenone has joined #pypy

05:25 pilne has quit [Quit: Quitting!]

05:41 <bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1580 [default]

05:41 forgottenone has quit [Quit: Konversation terminated!]

05:49 __Yiota has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

06:37 hvariant has quit [Remote host closed the connection]

06:37 zmt00 has quit [Quit: Leaving]

06:48 <bbot2> Failure: http://buildbot.pypy.org/builders/jitbackendonly-own-linux-armhf/builds/1635

07:46 drolando has quit [Remote host closed the connection]

07:47 drolando has joined #pypy

08:10 realitix has joined #pypy

08:12 oberstet has joined #pypy

08:29 <bbot2> Success: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/3162

08:31 raynold has quit [Quit: Connection closed for inactivity]

08:40 antocuni has joined #pypy

08:51 Rotonen has quit [Ping timeout: 240 seconds]

08:57 Rotonen has joined #pypy

09:05 cstratak has joined #pypy

09:15 marr has joined #pypy

09:34 __Yiota has joined #pypy

10:04 _stian has joined #pypy

10:07 <_stian> Hey, is it possible to make similar rewrite rules that INT have to LongObjects (like mul and div to shifts)?

10:11 <cfbolz> in the JIT? not easily

10:11 <cfbolz> _stian: that would rewrite calls, which we don't have much support for right now

10:11 <_stian> Ok

10:13 <cfbolz> _stian: you are thinking of something like implementing longnum * 2 by longnum << 1?

10:13 <_stian> The case is handled in rbigint, but calling right/lshift directly is still way faster.

10:13 <_stian> Correct

10:19 antocuni has quit [Ping timeout: 240 seconds]

10:21 <_stian> Hmm, maybe not as much as I remember, it only makes 2% difference.

10:22 <cfbolz> _stian: where is the code in rbigint that handles it?

10:23 <cfbolz> ah, one of the cases in int_mul, I suppose

10:23 <_stian> 2% after the JIT kicks in it seems, doing a loop too few times to be optimized seem to have a 50% cost.

10:23 <_stian> ye

10:24 <_stian> digit & (digit - 1) == 0: result = self.lqshift(ptwotable[digit])

10:26 <cfbolz> if it becomes super important, there is still something you can do: you can carefully use @jit.unroll_iff(lambda self, b: jit.isconstant(b)) on int_mul to achieve some of what you describe

10:27 <cfbolz> but then int_mul would be no longer elidable, which is also not good

10:28 mattip has quit [Ping timeout: 260 seconds]

10:28 alex_ has joined #pypy

10:29 <_stian> Most of the difference seem to be on unjitted code, doing a loop 100 times with shift gives me 0.16s with mul vs 0.09s with shift, over 50%. While at 10.000 it gives 7.3s to 7.16s, about 2%.

10:29 mattip has joined #pypy

10:29 <_stian> So I guess it isn't that important.

10:31 traverseda has quit [Ping timeout: 248 seconds]

11:08 __Yiota has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

11:30 jcea has joined #pypy

11:40 antocuni has joined #pypy

11:46 __Yiota has joined #pypy

12:14 __Yiota has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

12:19 realitix has quit [Ping timeout: 246 seconds]

12:19 realitix has joined #pypy

12:21 __Yiota has joined #pypy

12:27 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3460 [default]

12:27 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3461 [py3.5]

12:39 adamholmberg has joined #pypy

12:40 <bbot2> Success: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/2370

13:07 <_stian> Have __slots__ any special meaning in rpython, or is it just used in python / uncompiled mode? Just tried adding it to a class to see, and the binary increased by about 100byte.

13:22 exarkun has quit [Ping timeout: 252 seconds]

13:22 lritter__ has quit [Remote host closed the connection]

13:25 adamholmberg has quit [Remote host closed the connection]

13:26 __Yiota has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]

13:26 adamholmberg has joined #pypy

13:26 nedbat has quit [Ping timeout: 255 seconds]

13:26 exarkun has joined #pypy

13:28 adamholmberg has quit [Remote host closed the connection]

13:29 adamholmberg has joined #pypy

13:29 nedbat has joined #pypy

13:34 __Yiota has joined #pypy

13:37 _stian has quit [Ping timeout: 260 seconds]

14:05 exarkun has quit [Ping timeout: 240 seconds]

14:11 exarkun has joined #pypy

14:18 forgottenone has joined #pypy

14:24 yuyichao has quit [Ping timeout: 248 seconds]

14:35 yuyichao has joined #pypy

14:46 marky1991 has joined #pypy

14:55 <cfbolz> _stian: slots have no meaning

14:55 <cfbolz> And the binary size can fluctuate a lot just by recompiling

15:11 exarkun has quit [Ping timeout: 246 seconds]

15:16 exarkun has joined #pypy

15:28 zmt00 has joined #pypy

15:29 jamesaxl has quit [Read error: Connection reset by peer]

15:32 jamesaxl has joined #pypy

15:33 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3461 [py3.5]

15:42 adamholm_ has joined #pypy

15:45 adamholmberg has quit [Ping timeout: 260 seconds]

15:56 exarkun has quit [Ping timeout: 255 seconds]

15:59 realitix has quit [Quit: Leaving]

16:00 exarkun has joined #pypy

16:02 WGH has quit [Ping timeout: 255 seconds]

16:02 tbodt has joined #pypy

16:04 WGH has joined #pypy

16:05 raynold has joined #pypy

16:05 marky1991 has quit [Remote host closed the connection]

16:06 tbodt has quit [Client Quit]

16:06 exarkun has quit [Ping timeout: 240 seconds]

16:06 tbodt has joined #pypy

16:07 marky1991 has joined #pypy

16:13 marky1991 has quit [Remote host closed the connection]

16:14 exarkun has joined #pypy

16:15 marky1991 has joined #pypy

16:15 marky1991 has quit [Remote host closed the connection]

16:16 marky1991 has joined #pypy

16:27 marky1991 has quit [Remote host closed the connection]

16:27 oberstet has quit [Ping timeout: 240 seconds]

16:28 marky1991 has joined #pypy

16:32 marky1991 has quit [Ping timeout: 248 seconds]

16:48 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

16:50 oberstet has joined #pypy

16:52 <mattip> arigato: any progress on reviewing the buildbot changes?

17:02 <ionelmc> would cffi questions be answered here?

17:03 <ionelmc> can i just pass a big lump of C code i wanna use in python to ffibuilder.set_source?

17:05 cstratak has quit [Quit: Leaving]

17:08 antocuni has quit [Ping timeout: 246 seconds]

17:10 antocuni has joined #pypy

17:10 tbodt has joined #pypy

17:11 tbodt has quit [Client Quit]

17:12 <antocuni> ionelmc: yes

17:14 <antocuni> ionelmc: you can also pass sources=['file1.c', 'file2.c']

17:14 <antocuni> they will be compiled and linked together with the extension

17:15 <antocuni> basically, most of the kwargs you pass are forwarded to distutils.Extension

17:15 tbodt has joined #pypy

17:15 tbodt has quit [Client Quit]

17:23 marky1991 has joined #pypy

17:38 antocuni has quit [Ping timeout: 240 seconds]

18:09 kipras`away is now known as kipras

18:12 tbodt has joined #pypy

18:18 OOPMan has joined #pypy

18:32 <OOPMan> I don't suppose anyone in here has a lot of experience with async generators and/or comprehensions from python 3.6?

18:32 <OOPMan> I don't suppose you know of any async friendly equivalent to `reduce`?

18:33 exarkun has quit [Read error: Connection reset by peer]

18:37 <nanonyme> OOPMan, I'd be wary of things that do a lot of context switching in any case, async programming or threading

18:40 exarkun has joined #pypy

18:43 <OOPMan> nanonyme: Not really answering my question but based on some googling it seems the answer is basically no

18:44 <OOPMan> Looks like it shouldn't be hard to implement though

18:44 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

18:44 OOPMan has quit [Quit: Konversation terminated!]

19:01 redj has quit [Ping timeout: 248 seconds]

19:05 tbodt has joined #pypy

19:09 <ionelmc> what did antocuni mean by sources=... ?

19:10 <ionelmc> set_sources(..., sources=[...]) i guess?

19:10 marky1991 has quit [Ping timeout: 255 seconds]

19:22 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

19:25 tbodt has joined #pypy

19:26 tbodt has quit [Client Quit]

19:28 tbodt has joined #pypy

19:38 antocuni has joined #pypy

19:40 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

19:44 tbodt has joined #pypy

19:50 tbodt has quit [Ping timeout: 255 seconds]

19:51 tbodt has joined #pypy

19:54 redj has joined #pypy

19:57 antocuni has quit [Ping timeout: 240 seconds]

19:58 rubdos has quit [Ping timeout: 255 seconds]

20:01 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

20:06 tbodt has joined #pypy

20:13 <ronan> pff... _ctypes is a mess

20:14 <exarkun> heh heh heh.

20:16 marky1991 has joined #pypy

20:17 forgottenone has quit [Quit: Konversation terminated!]

20:23 zmt01 has joined #pypy

20:26 zmt00 has quit [Ping timeout: 240 seconds]

20:30 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

20:49 antocuni has joined #pypy

20:57 tbodt has joined #pypy

21:03 marky1991 has quit [Ping timeout: 255 seconds]

21:07 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

21:15 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-armhf-raspbian/builds/1450

21:24 tbodt has joined #pypy

21:26 tbodt has quit [Client Quit]

21:29 tbodt has joined #pypy

21:35 tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]

21:35 marky1991 has joined #pypy

21:38 <kenaan> stian math-improvements 2a90db93be84 /rpython/rlib/rbigint.py: replace construction of a onedigit with premade, and mark sign and size as immutable (which is true post ...

21:38 <kenaan> stian math-improvements 7427553eee6b /rpython/rlib/rbigint.py: Digits don't change post-construction

21:38 <kenaan> stian math-improvements e80b5cdab0a6 /rpython/rlib/: Fix int_mod with mod argument

21:56 oberstet has quit [Ping timeout: 240 seconds]

21:58 tbodt has joined #pypy

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-s390x/builds/744 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4288 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5018 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/6318 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/5526 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1581 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3462 [default]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/jitbackendonly-own-linux-armhf/builds/1636