00:19
andi- has quit [Ping timeout: 248 seconds]
00:25
andi- has joined #pypy
00:26
<
jacob22 >
lesshaste: You may have success with the C API of gmpy2.
00:29
igitoor has quit [Ping timeout: 260 seconds]
00:31
igitoor has joined #pypy
00:40
xcm has quit [Remote host closed the connection]
00:41
xcm has joined #pypy
00:45
igitoor has quit [Changing host]
00:45
igitoor has joined #pypy
01:11
lritter has quit [Ping timeout: 268 seconds]
01:30
jvesely has joined #pypy
01:59
jcea has quit [Quit: jcea]
02:28
LarstiQ has quit [Ping timeout: 240 seconds]
02:28
LarstiQ has joined #pypy
02:42
jcea has joined #pypy
02:51
dddddd has quit [Remote host closed the connection]
03:09
jcea has quit [Remote host closed the connection]
04:37
xcm has quit [Remote host closed the connection]
04:39
xcm has joined #pypy
04:39
zmt00 has joined #pypy
04:42
zmt01 has quit [Ping timeout: 260 seconds]
05:16
riddle has quit [Ping timeout: 265 seconds]
06:35
jvesely has quit [Quit: jvesely]
06:54
riddle has joined #pypy
08:21
<
lesshaste >
cfbolz, thanks
08:21
<
lesshaste >
jacob22, good point
08:52
<
lesshaste >
I have a very weird slowdown in pypy3 if I just increase the range of integers from 0...2**63-1 to 0..2**64-1. Is there a good tool to profile the code to tell which line is doing it?
08:57
<
lesshaste >
if you just change the 63 to 64 it slows down by a factor of 4
08:57
<
lesshaste >
why is that?
09:07
<
Dejan >
maybe because the difference between 2^64 and 2^63 is 8*10^24 :)
09:10
<
Dejan >
while 2^63-2^62 = 4*10^24 (4 times less)
09:19
xcm has quit [Remote host closed the connection]
09:20
xcm has joined #pypy
09:24
<
cfbolz >
lesshaste: 2**63-1 is the biggest number representable by a machine long
09:24
<
cfbolz >
So going beyond that is slower
09:33
<
lesshaste >
cfbolz, oh because it is signed?
09:33
<
cfbolz >
lesshaste: yep
09:34
<
lesshaste >
cfbolz, hmm... does this mean that popcount won't be safe to rely on too?
09:34
<
lesshaste >
I am computing the hamming distance between two binary arrays by representing them as ints, doing XOR and then popcount
09:35
<
lesshaste >
A factor of 4 slow down does still seem odd to me as there is no multiplication going on. Just XOR and popcount
09:37
<
lesshaste >
Dejan, :)
09:39
<
cfbolz >
lesshaste: if you leave the range, the next step is using two integers for your values, so that's 2x right there
09:39
<
lesshaste >
cfbolz, yes
09:39
<
lesshaste >
cfbolz, I was wondering where the other 2x comes from
09:39
<
cfbolz >
Everything is suddenly a software algorithm
09:39
<
cfbolz >
lesshaste: if you want to share your code, we could take a look
09:40
<
lesshaste >
cfbolz, ^^
09:41
<
cfbolz >
There should really be a popcount in python
09:42
<
lesshaste >
yes. I was looking at julia and they have count_ones
09:48
<
lesshaste >
I do wonder if I should just move over to julia but I do like a popular language and there is a chance julia will never be popular
09:59
<
Dejan >
so you are a hipster
09:59
<
Dejan >
well, Python devs are hipsters too :)
10:00
<
Dejan >
my approach is - use the best tool for the job
10:17
<
lesshaste >
Dejan, true but for scientific research Python is awesome because of ease of use and the massive number of libraries and terrible because it is as slow as a snail :)
10:19
<
lesshaste >
Dejan, the lack of help to spot typos is also a serious problem
11:02
msjyoo has joined #pypy
11:28
dddddd has joined #pypy
12:04
speeder39_ has joined #pypy
12:38
<
phlebas >
lesshaste: re spotting typos; i was pleasantly surprised how well the microsoft language server for python works (even for the pypy codebase). works nicely in emacs and vi, not just vscode and such
12:43
<
cfbolz >
phlebas: ah, cool. at some point you need to tell me about it, I never managed to set it up nice for vscode
13:11
jcea has joined #pypy
13:15
jcea has quit [Ping timeout: 245 seconds]
13:35
jacob22 has quit [Read error: Connection reset by peer]
13:38
jacob22 has joined #pypy
14:09
Rhy0lite has joined #pypy
14:18
jvesely has joined #pypy
14:55
<
mattip >
finally got to a point where py3.6 to default benchmark comparisons are useful. There is alot of historical cruft that gets in the way
14:56
<
mattip >
for instance, here is a direct comparison of a py3.6 run to a default run
14:59
<
mattip >
spitfire_cstring2 is 22x slower, I wonder if that is true or a glitch
15:00
<
mattip >
django, telco are slower by 3x, 4x
15:02
jcea has joined #pypy
15:06
jcea has quit [Remote host closed the connection]
15:06
jcea has joined #pypy
15:22
Smigwell has joined #pypy
15:25
<
mattip >
simplified link
15:26
DarkPlutonium is now known as DarkUranium
15:30
YannickJadoul has joined #pypy
15:31
<
Dejan >
we all go back to Python2 now
15:32
<
Dejan >
there are no cases when Python2 is considerably slower
15:35
<
mattip >
the benchmarks started as python2 code, so I would be surprised if python3 was faster
15:36
<
Dejan >
There are quite few where python3 is tinsy bit faster
15:39
<
mattip >
anything around 5% is probably just noise
15:39
<
jacob22 >
There ought to be some low hanging fruit, now that we have actual measurements.
15:41
<
mattip >
I think I am pretty much happy with the way speed.pypy.org is working now, if there is something missing let me know
15:41
<
Dejan >
mattip, it is awesome
15:43
<
jacob22 >
Comparisons of CPython3 and PyPy3 are what end users will ultimately be most interested in.
15:44
<
mattip >
we could graph that on the home page: either as a separate graph or as part of the 2nd graph in a different color
15:45
<
mattip >
but that "summary" is kind of misleading
15:46
<
mattip >
I guess it would be sort of easy to do a cpython3 benchmark run, and then the comparison page would have the info
15:51
<
cfbolz >
mattip: thanks for much for doing this!
16:03
speeder39_ has quit [Quit: Connection closed for inactivity]
16:07
lritter has joined #pypy
16:13
jvesely has quit [Quit: jvesely]
16:50
string has joined #pypy
17:14
jacob22 has quit [Quit: Konversation terminated!]
17:15
<
lesshaste >
phlebas, I don't think I know about that!
17:29
rubdos has quit [Quit: WeeChat 2.4]
17:34
rubdos has joined #pypy
17:35
rubdos has quit [Client Quit]
17:39
rubdos has joined #pypy
17:40
rubdos has quit [Client Quit]
17:55
rubdos has joined #pypy
17:59
YannickJadoul has quit [Quit: Leaving]
18:06
fryguybob has quit [Ping timeout: 260 seconds]
18:06
fryguybob has joined #pypy
18:32
rubdos has quit [Quit: WeeChat 2.4]
18:39
rubdos has joined #pypy
18:42
jvesely has joined #pypy
19:30
jacob22 has joined #pypy
19:46
rubdos has quit [Quit: WeeChat 2.4]
19:54
rubdos has joined #pypy
20:21
zmt00 has quit [Quit: Leaving]
20:21
zmt00 has joined #pypy
20:27
Rhy0lite has quit [Quit: Leaving]
20:51
igitoor has quit [Ping timeout: 272 seconds]
20:52
igitoor has joined #pypy
20:55
<
kenaan >
mattip py3.6 95ba888acdd6 /pypy/module/cpyext/include/Python.h: issue 3160: include structseq.h (needed for PyStructSequence_InitType2 in NumPy)
21:00
igitoor has quit [Changing host]
21:00
igitoor has joined #pypy
21:23
Smigwell has left #pypy [#pypy]
21:57
adamholmberg has joined #pypy
22:08
jacob22 has quit [Quit: Konversation terminated!]
22:09
jacob22 has joined #pypy
22:31
jacob22 has quit [Ping timeout: 265 seconds]
22:53
Ai9zO5AP has quit [Ping timeout: 240 seconds]
22:55
Ai9zO5AP has joined #pypy
23:38
jacob22 has joined #pypy