01:53
anushi has quit [Quit: Bye]
01:55
jaafar has joined #ste||ar
02:41
K-ballo has quit [Quit: K-ballo]
02:44
hkaiser has quit [Quit: bye]
03:10
bita has joined #ste||ar
03:58
bita has quit [Read error: Connection reset by peer]
06:17
ste||ar-github has joined #ste||ar
06:17
<
ste||ar-github >
hpx/master b4bc3b7 John Biddiscombe: Merge pull request #3457 from STEllAR-GROUP/fixing_warnings...
06:17
ste||ar-github has left #ste||ar [#ste||ar]
06:26
jaafar has quit [Ping timeout: 244 seconds]
07:22
daissgr has joined #ste||ar
07:31
david_pfander has joined #ste||ar
08:19
david_pfander has quit [Quit: david_pfander]
08:21
david_pfander has joined #ste||ar
08:21 _diers_ has quit [Quit: _diers_]
08:54
mcopik_ has quit [Ping timeout: 245 seconds]
09:28
ste||ar-github has joined #ste||ar
09:28
<
ste||ar-github >
hpx/master 9820039 Mikael Simberg: Add quickstart examples as tests...
09:28
<
ste||ar-github >
hpx/master 6df436b Mikael Simberg: Merge pull request #3451 from msimberg/test-examples...
09:28
ste||ar-github has left #ste||ar [#ste||ar]
09:52
Anushi1998 has joined #ste||ar
09:58
mcopik_ has joined #ste||ar
10:02
ste||ar-github has joined #ste||ar
10:02
ste||ar-github has left #ste||ar [#ste||ar]
10:07
ste||ar-github has joined #ste||ar
10:07
<
ste||ar-github >
hpx/gh-pages 5e0a4b8 StellarBot: Updating docs
10:07
ste||ar-github has left #ste||ar [#ste||ar]
10:27
K-ballo has joined #ste||ar
11:27
hkaiser has joined #ste||ar
11:48
mcopik_ has quit [Ping timeout: 252 seconds]
12:01
mcopik_ has joined #ste||ar
12:08
Anushi1998 has quit [Quit: Bye]
13:05
hkaiser has quit [Quit: bye]
13:08
eschnett has quit [Quit: eschnett]
13:15
aserio has joined #ste||ar
13:31
eschnett has joined #ste||ar
13:39 _diers_ has joined #ste||ar
13:44
hkaiser has joined #ste||ar
13:50
mcopik_ has quit [Ping timeout: 250 seconds]
13:52
mcopik_ has joined #ste||ar
13:55
quaz0r has quit [Ping timeout: 272 seconds]
13:56
quaz0r has joined #ste||ar
14:03
mcopik_ has quit [Ping timeout: 246 seconds]
14:04
mcopik_ has joined #ste||ar
15:16
mcopik_ has quit [Ping timeout: 240 seconds]
15:24
jaafar has joined #ste||ar
15:32
dmarce1 has joined #ste||ar
15:46
mcopik_ has joined #ste||ar
16:33
mcopik_ has quit [Ping timeout: 250 seconds]
16:42
<
nikunj97 >
hkaiser, see pm please
16:46
mcopik_ has joined #ste||ar
17:19
ste||ar-github has joined #ste||ar
17:19
<
ste||ar-github >
hpx/thread_local_allocator c451672 Hartmut Kaiser: Adding thread local allocator and use it for future shared states
17:19
<
ste||ar-github >
hpx/thread_local_allocator 95fa0f6 Hartmut Kaiser: Removing the need for one of the atomics
17:19
<
ste||ar-github >
hpx/thread_local_allocator 7fc9e35 Hartmut Kaiser: Cleaning up thread_allocator...
17:19
ste||ar-github has left #ste||ar [#ste||ar]
17:19
<
hkaiser >
heller: I updated the thread allocator - I think I addressed everything we talked about
17:19
<
hkaiser >
a couple of flyby optimizations are on that branch as well
17:38
aserio has quit [Ping timeout: 240 seconds]
17:50
mcopik_ has quit [Ping timeout: 245 seconds]
17:56
aserio has joined #ste||ar
17:56
<
heller >
hkaiser: cool! I'll have a look!
17:56
<
heller >
hkaiser: did you get my mail?
17:59
<
heller >
hkaiser: did you measure the improvements?
18:01
<
zao >
You people seem to have been busy lately, judging by the commit stream.
18:06
Vir has quit [Ping timeout: 240 seconds]
18:07
Vir has joined #ste||ar
18:15
daissgr has quit [Ping timeout: 250 seconds]
18:15
<
heller >
hkaiser: I'm rerunning the benchmarks used in my dissertation against your branch
18:23
<
heller >
hkaiser: mind if I push the fixes for gcc?
18:28
<
K-ballo >
what could possibly go wrong?
18:29
<
heller >
for his PR, not master ;)
18:31
* K-ballo
isn't following along
18:31
<
K-ballo >
heller: I still can't adjust to your new avatar!
18:31
<
heller >
guess i'll revert it
18:32
<
heller >
I'm not sure either ;)
18:37
ste||ar-github has joined #ste||ar
18:37
ste||ar-github has left #ste||ar [#ste||ar]
18:41
<
heller >
hkaiser: I'm getting lot's of errors in distributed
18:41
<
heller >
hkaiser: {what}: null thread id encountered (is this executed on a HPX-thread?): HPX(null_thread_id)
18:42
<
heller >
hkaiser: src/runtime/threads/thread_data.cpp line 111
18:43
<
heller >
all coming up at pre main
18:55
eschnett has quit [Quit: eschnett]
18:57
<
heller >
some kind of strange bug ... doesn't seem to happen in a debug build
18:58
<
heller >
running a ubsan build now... this is reproducable with hello_world
19:02
mcopik_ has joined #ste||ar
19:08
<
nikunj97 >
heller, is it related to hpx_main?
19:08
<
hkaiser >
heller: yah, I know - fixing that right now
19:08
hkaiser has quit [Quit: bye]
19:11
eschnett has joined #ste||ar
19:13
<
heller >
I'm also not fond of that union "cast"
19:15
<
K-ballo >
union cast? I can't allow that...
19:27
<
K-ballo >
where was that lifted from?
19:29
daissgr has joined #ste||ar
19:38
hkaiser has joined #ste||ar
20:01
aserio1 has joined #ste||ar
20:02
aserio has quit [Ping timeout: 240 seconds]
20:02
aserio1 is now known as aserio
20:19
khuck has joined #ste||ar
20:20
<
khuck >
bibek: who is the lead author of the Phylanx ESPM2 paper?
20:21
<
khuck >
bibek: I am curious if comparing Phylanx to Theano is a good idea - it's a dead project now
20:22
<
hkaiser >
khuck: Rod is the first author
20:23
<
khuck >
ok, - just curious. see above comment about Theano
20:26
<
hkaiser >
heller: fixed now
20:26
<
hkaiser >
thanks for your patience
20:26
ste||ar-github has joined #ste||ar
20:26
<
ste||ar-github >
hpx/thread_local_allocator 418ad7a Hartmut Kaiser: Partially roll back some of the previous changes
20:26
ste||ar-github has left #ste||ar [#ste||ar]
20:26
<
hkaiser >
khuck: yah, good point
20:28
<
khuck >
of course, the roadside is littered with the bones of those that have come before... :)
20:37
<
hkaiser >
heller: heh, any other way?
20:46
aserio has quit [Ping timeout: 252 seconds]
20:46
dmarce1 has quit [Quit: Leaving.]
20:48
<
K-ballo >
both alternatives are integers, so that's just bit twidling, isn't it?
20:49
<
K-ballo >
it's not even a fancy union cast
20:49
<
K-ballo >
where was that lifted from?
21:00
eschnett has quit [Quit: eschnett]
21:07
<
hkaiser >
K-ballo: sure it's bit-fiddling
21:07
<
hkaiser >
you mean I should just mask and shift?
21:10
<
K-ballo >
it has the advantage of being defined.. is there a reason not to?
21:11
<
hkaiser >
no reason, wasn't thinking about this too much
21:19
ste||ar-github has joined #ste||ar
21:19
<
ste||ar-github >
hpx/thread_local_allocator 4abed31 Hartmut Kaiser: Partially roll back some of the previous changes
21:19
<
ste||ar-github >
hpx/thread_local_allocator 6bec6ba Hartmut Kaiser: Replace union cast with bit twiddling
21:19
ste||ar-github has left #ste||ar [#ste||ar]
21:19
<
hkaiser >
heller, K-ballo: better now ^^ ?
21:21
aserio has joined #ste||ar
21:22
ste||ar-github has joined #ste||ar
21:22
<
ste||ar-github >
hpx/thread_local_allocator ed68ca1 Hartmut Kaiser: Remove unneeded volatile keywords
21:22
ste||ar-github has left #ste||ar [#ste||ar]
21:34
aserio has quit [Quit: aserio]
21:40
parsa[w] has quit [Ping timeout: 250 seconds]
21:43
parsa[w] has joined #ste||ar
21:43
bibek has quit [Quit: Konversation terminated!]
22:03
<
khuck >
hkaiser: what's the status on the paper? I just pushed a bunch of "todo" items and revisions.
22:20
<
hkaiser >
khuck: no idea, Rod/Bibek are on top of this
22:20
<
hkaiser >
I read over everything and tweaked some minor things
22:30
<
khuck >
hkaiser: I think the related work needs a lot of work. And there are some other corrections that aren't trivial
22:44
daissgr has quit [Quit: WeeChat 1.9.1]
23:28
<
heller >
hkaiser: thanks ;)
23:32
<
heller >
hkaiser: also, are we sure the tagged ptr technique works on all our platforms?
23:35
<
heller >
ah yes, we are since we have the addresses nicely aligned on page boundaries
23:39
<
heller >
hkaiser: another thing, reinterpret_cast isn't constexpr
23:49
<
heller >
rerunning the benchmarks now
23:50
<
heller >
distributed looks better now, thanks
23:51
<
hkaiser >
heller: remember you reminded me that I had implemented local-only id-types
23:51
<
hkaiser >
everything was in place, just the resolv eoperation was missing ;-)
23:54
<
heller >
did you implement it now?
23:54
<
hkaiser >
yah, it's on that branch
23:58
<
heller >
how much did it improve the situation?