00:08
adamholmberg has quit [Remote host closed the connection]
00:09
adamholmberg has joined #pypy
00:13
adamholmberg has quit [Ping timeout: 240 seconds]
00:20
exarkun has quit [Ping timeout: 248 seconds]
00:21
antocuni has quit [Ping timeout: 260 seconds]
00:25
exarkun has joined #pypy
00:58
yuyichao_ has quit [Ping timeout: 268 seconds]
01:11
yuyichao_ has joined #pypy
01:16
marr has quit [Ping timeout: 255 seconds]
01:50
marky1991 has joined #pypy
02:12
songww has joined #pypy
02:32
slacky has quit [Ping timeout: 248 seconds]
02:42
jcea has quit [Quit: jcea]
02:45
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
02:53
pilne has quit [Quit: Quitting!]
03:30
ArneBab_ has joined #pypy
03:30
ArneBab_ has quit [Changing host]
03:30
ArneBab_ has joined #pypy
03:35
ArneBab has quit [Ping timeout: 264 seconds]
03:41
slacky has joined #pypy
03:47
slacky has quit [Ping timeout: 248 seconds]
04:08
kushal has quit [Ping timeout: 240 seconds]
04:09
kushal has joined #pypy
04:09
kushal is now known as Guest78064
04:41
slacky has joined #pypy
04:48
slacky has quit [Ping timeout: 240 seconds]
05:18
jamesaxl has joined #pypy
05:25
Nizumzen has joined #pypy
05:29
astronavt has joined #pypy
05:41
slacky has joined #pypy
05:47
slacky has quit [Ping timeout: 276 seconds]
06:09
zmt00 has quit [Quit: Leaving]
06:41
slacky has joined #pypy
06:46
slacky has quit [Ping timeout: 248 seconds]
07:32
songww_ has joined #pypy
07:36
songww has quit [Ping timeout: 268 seconds]
07:36
songww_ is now known as songww
07:41
slacky has joined #pypy
07:47
slacky has quit [Ping timeout: 276 seconds]
07:48
hubert0 has joined #pypy
07:48
hubert0 has quit [Remote host closed the connection]
07:48
hubert0 has joined #pypy
07:59
marky1991 has quit [Ping timeout: 268 seconds]
08:10
<
kenaan >
mattip win32-vcvars 7f64b9246539 /rpython/translator/platform/windows.py: maybe fix?
08:14
songww has quit [Remote host closed the connection]
08:15
songww has joined #pypy
08:19
Nizumzen has quit [Ping timeout: 240 seconds]
08:22
<
kenaan >
mattip win32-vcvars 5d23987c65b3 /rpython/translator/platform/windows.py: typo
08:41
slacky has joined #pypy
08:42
<
kenaan >
mattip win32-vcvars b6119a8ea747 /pypy/doc/whatsnew-head.rst: both build slaves seem happy
08:42
<
kenaan >
mattip default d1aaa6aca19d /: merge win32-vcvars, log more and try vsvars32 before vcvars32, go figure
08:44
<
mattip >
all this vcvars crap is so broken
08:47
slacky has quit [Ping timeout: 240 seconds]
08:58
marr has joined #pypy
08:58
mattip has left #pypy ["bye"]
09:02
Nizumzen has joined #pypy
09:06
Nizumzen has quit [Client Quit]
09:12
Nizumzen has joined #pypy
09:26
Joannah has joined #pypy
09:41
slacky has joined #pypy
09:45
Joannah has quit [Ping timeout: 260 seconds]
09:46
antocuni has joined #pypy
09:47
slacky has quit [Ping timeout: 248 seconds]
10:41
slacky has joined #pypy
10:47
slacky has quit [Ping timeout: 255 seconds]
11:23
antocuni has quit [Ping timeout: 248 seconds]
11:39
slacky has joined #pypy
11:51
astronavt has quit [Remote host closed the connection]
12:08
jcea has joined #pypy
12:40
raynold has quit [Quit: Connection closed for inactivity]
12:54
songww has quit [Read error: Connection reset by peer]
12:54
songww has joined #pypy
12:56
songww has quit [Remote host closed the connection]
13:13
Rhy0lite has joined #pypy
13:26
antocuni has joined #pypy
13:33
<
antocuni >
I'm trying to understand how much time gc-collect-step takes
13:34
<
antocuni >
is there any standard and/or blessed way to convert PYPYLOG timestamps into seconds?
13:50
<
fijal >
antocuni: look into logparser
13:51
<
fijal >
It'll show you in %not seconds
13:51
<
fijal >
(it's a tick which means you divide by CPU frequency)
13:52
mattip has joined #pypy
13:53
<
antocuni >
so to know the actual seconds I need to know the frequency?
13:54
<
cfbolz >
you can't. it's changing too much
13:58
<
mattip >
kind of stuck in numpy. I got the updateifcopy flag refactored out of the ndarray dealloc, but was left with what to do with nditer
13:58
<
antocuni >
"Recent Intel processors include a constant rate TSC (identified by the kern.timecounter.invariant_tsc sysctl on FreeBSD or by the "constant_tsc" flag in Linux's /proc/cpuinfo). With these processors, the TSC ticks at the processor's nominal frequency, regardless of the actual CPU clock frequency"
13:58
<
mattip >
specifically what happens if an nditer creates operands with updateifcopy semantics
13:59
<
mattip >
so I thought, OK, context manager and started a pull request
13:59
<
antocuni >
so in theory, I should just divide by the nominal frequency of my CPU
13:59
<
mattip >
but the operands leak out
14:00
<
mattip >
after closing the iterator, the operand and the view are still "alive" but invalid
14:02
<
mattip >
much like a file or socket after closing. But an ndarray has no "open/closed" status
14:02
<
mattip >
any bright ideas?
14:05
<
antocuni >
cfbolz: indeed, it seems pretty accurate: I tried to measure both a time.sleep(10) and a busy loop of 10 seconds, so that in theory the CPU should scale the frequency
14:06
<
antocuni >
and I get more or less the same value (99.6% of accuracy), which is 2.7Ghz*10
14:06
<
antocuni >
good to know, I suppose :)
14:23
<
cfbolz >
antocuni: cool
14:28
marky1991 has joined #pypy
14:29
marky1991 has quit [Remote host closed the connection]
14:29
marky1991 has joined #pypy
14:34
<
kenaan >
cfbolz default a805f563cfd0 /rpython/jit/metainterp/optimizeopt/: - hypothesis tests for the IntBound methods - lighgly refactor optimization of some int_ ops to make hypothesis te...
14:38
adamholmberg has joined #pypy
14:38
<
cfbolz >
mattip: in cpython this all depends on refcounts?
14:40
<
mattip >
correct. NumPy has this concept of a scratch buffer that can be attached to a ndarray, and the condition to write the temp data from the scratch buffer back to the main data was dealloc
14:40
<
mattip >
now I made a function to be called just before dealloc that does the writeback
14:41
<
mattip >
so you have to keeep track of "the last known ref to the ndarray with scratch buffer"
14:41
<
mattip >
the concept only leaks out of C level numpy in nditer
14:42
<
mattip >
if you have time you can read the whole history of the nditer saga here
14:43
<
mattip >
but that pastebin above summarizes the situation best (refactor of the last comment in the pull request)
14:44
tormoz has quit [Remote host closed the connection]
14:47
tormoz has joined #pypy
14:48
marky1991 has quit [Remote host closed the connection]
14:49
<
cfbolz >
mattip: this all sounds very annoying indeed :-(
14:50
<
mattip >
the most annoying thing is nditer is a very rarely used corner of numpy, with an enormous api
14:50
<
mattip >
IMO the best thing to do would be to deprecate most of it, if not all
14:51
<
mattip >
but that is impractical
14:52
hubert0 has quit [Remote host closed the connection]
14:52
<
mattip >
kind of like trying to deprecate micronumpy, the only way to find out if someone cares is to remove it
14:52
hubert0 has joined #pypy
14:53
tormoz has quit [Remote host closed the connection]
14:55
* cfbolz
-> teaching
14:56
tormoz has joined #pypy
14:57
hubert0 has quit [Ping timeout: 252 seconds]
14:57
marky1991 has joined #pypy
15:00
mattip has left #pypy ["bye"]
15:01
yuyichao_ has quit [Ping timeout: 276 seconds]
15:11
yuyichao_ has joined #pypy
15:13
adamholmberg has quit [Read error: No route to host]
15:13
adamholmberg has joined #pypy
15:20
<
kenaan >
antocuni fix-vmprof-stacklet-switch-2 37b5a39510e7 /pypy/module/_continuation/test/test_stacklet.py: skip this when run with -A
15:20
<
kenaan >
antocuni fix-vmprof-stacklet-switch-2 4776dc8a84f7 /pypy/module/_continuation/test/test_translated.py: simplify
15:20
<
kenaan >
antocuni fix-vmprof-stacklet-switch-2 9ac249e058e3 /: add a test which checks that vmprof is actually enabled inside greenlets; it fails on defau...
15:20
<
kenaan >
antocuni fix-vmprof-stacklet-switch-2 f7fa5a2b8f51 /pypy/doc/whatsnew-head.rst: document this branch
15:22
<
fijal >
cfbolz, antocuni: if you're running at 100% CPU it should really be the nominal cpu
15:22
<
fijal >
Regardless of settings
15:23
<
antocuni >
fijal: yes, but the nice thing I discovered is that on "modern" CPUs, it is always incremented at the nominal frequency
15:24
<
cfbolz >
fijal: no, it's really not that simple any more
15:24
<
cfbolz >
But yes, seems you can get useful numbers anyway
16:04
nunatak has joined #pypy
16:10
mattip has joined #pypy
16:47
demonimin has quit [Remote host closed the connection]
16:51
demonimin has joined #pypy
17:01
tbodt has joined #pypy
17:01
zmt00 has joined #pypy
17:21
Nizumzen has quit [Ping timeout: 240 seconds]
17:29
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
17:33
drolando has joined #pypy
17:53
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
17:53
Nizumzen has joined #pypy
17:57
tbodt has joined #pypy
18:04
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
18:15
Guest78064 is now known as kushal
18:15
kushal has quit [Changing host]
18:15
kushal has joined #pypy
18:16
antocuni has quit [Ping timeout: 248 seconds]
18:25
marr has quit [Remote host closed the connection]
18:26
marky1991 has quit [Ping timeout: 252 seconds]
18:34
<
kenaan >
rlamy default b2e4b128808e /pypy/module/_io/: Improve interp-level test and fix bugs in W_TextIOWrapper.readline_w()
18:35
drolando has joined #pypy
18:35
drolando has quit [Remote host closed the connection]
18:35
drolando has joined #pypy
18:36
<
kenaan >
rlamy py3.5 fcf5d3fb56f4 /: hg merge default
18:36
marky1991 has joined #pypy
18:36
marky1991 has quit [Remote host closed the connection]
18:36
marky1991 has joined #pypy
18:57
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
18:59
tbodt has joined #pypy
19:03
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
19:05
marky1991 has quit [Ping timeout: 248 seconds]
19:17
drolando has joined #pypy
19:19
marky1991 has joined #pypy
19:19
marky1991 has quit [Changing host]
19:19
marky1991 has joined #pypy
19:20
marky1991 has quit [Remote host closed the connection]
19:20
marky1991 has joined #pypy
19:31
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
19:33
drolando has joined #pypy
19:51
antocuni has joined #pypy
19:54
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
19:56
tbodt has joined #pypy
20:03
kanaka has joined #pypy
20:03
kanaka has quit [Changing host]
20:03
kanaka has joined #pypy
20:11
raynold has joined #pypy
20:28
Rhy0lite has quit [Quit: Leaving]
20:30
hubert0 has joined #pypy
20:31
marr has joined #pypy
20:36
jamesaxl has quit [Quit: WeeChat 1.9.1]
20:59
hubert0 has quit [Remote host closed the connection]
20:59
hubert0 has joined #pypy
21:03
hubert0 has quit [Ping timeout: 240 seconds]
21:29
astronavt has joined #pypy
21:36
mattip has left #pypy ["bye"]
21:50
astronavt has quit [Remote host closed the connection]
21:50
astronavt has joined #pypy
21:51
astronavt has quit [Remote host closed the connection]
21:55
marky1991 has quit [Ping timeout: 252 seconds]
21:59
nunatak has quit [Quit: Leaving]
21:59
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
22:03
drolando has joined #pypy
22:03
drolando has quit [Remote host closed the connection]
22:04
drolando has joined #pypy
22:07
marky1991 has joined #pypy
22:36
drolando has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
22:36
adamholmberg has quit [Remote host closed the connection]
22:36
adamholmberg has joined #pypy
22:38
drolando has joined #pypy
22:38
drolando has quit [Remote host closed the connection]
22:39
drolando has joined #pypy
22:47
<
runciter >
can i help make that happen?
22:52
adamholmberg has quit [Remote host closed the connection]
22:52
adamholmberg has joined #pypy
22:53
adamholmberg has quit [Remote host closed the connection]
22:53
adamholmberg has joined #pypy
22:57
<
fijal >
tumbleweed: ^^^
22:57
<
fijal >
runciter: that's the person to pester
22:58
<
runciter >
fijal: ty!
22:58
<
runciter >
tumbleweed: let me know if i can help update the PPA to PyPy 5.9
23:05
dcrosta has joined #pypy
23:15
marky1991 has quit [Ping timeout: 248 seconds]
23:19
marky1991 has joined #pypy
23:32
marky1991 has quit [Remote host closed the connection]
23:33
marky1991 has joined #pypy
23:42
adamholmberg has quit [Remote host closed the connection]
23:43
adamholmberg has joined #pypy
23:47
adamholmberg has quit [Ping timeout: 255 seconds]
23:58
pilne has joined #pypy