aserio changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar.cct.lsu.edu | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | Buildbot: http://rostam.cct.lsu.edu/ | Log: http://irclog.cct.lsu.edu/
EverYoung has quit [Remote host closed the connection]
<github>
[hpx] biddisco closed pull request #2922: This fixes CircleCI errors when running with --hpx:bind=none (master...fixing_bind_none) https://git.io/vdncs
<github>
[hpx] biddisco force-pushed fixing_2914 from 1e758f4 to 2e59f03: https://git.io/vdnqV
<github>
hpx/fixing_2914 8b56042 John Biddiscombe: Custom pool executor was missing priority and stacksize options
<github>
hpx/fixing_2914 2e59f03 John Biddiscombe: Rename customized_pool_executor to pool_executor
<github>
[hpx] biddisco force-pushed alloc_membind from afb5e6b to c14fa15: https://git.io/vdnl9
<github>
hpx/alloc_membind c14fa15 John Biddiscombe: Remove extra #include and fix inspect problems
kisaacs has joined #ste||ar
<jbjnr>
pree: I made 2 PRs yesterday and a third is in preparation. I think I'll get my 4 by the end of the month. (I liked last year's T-shirt, so am doing it again to get another)
<jbjnr>
(But to be honest - it's pure coincidence that these PR's just happened to happen in Octover.)
<pree>
jbjnr : did you sign in ?
<jbjnr>
yes
<jbjnr>
last week
<pree>
jbjnr : congo : )
<pree>
okay
<jbjnr>
I do not know what "congo" means!
<pree>
*congrats
<pree>
congratulations
kisaacs has quit [Ping timeout: 240 seconds]
pree has quit [Ping timeout: 258 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
pree has joined #ste||ar
jaafar_ has quit [Ping timeout: 258 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
kisaacs has joined #ste||ar
zbyerly_ has quit [Ping timeout: 264 seconds]
kisaacs has quit [Ping timeout: 248 seconds]
<jbjnr>
oh dear, the green is all gone.
david_pfander has joined #ste||ar
Matombo has joined #ste||ar
dj9 has quit [Ping timeout: 260 seconds]
simbergm has joined #ste||ar
Matombo has quit [Remote host closed the connection]
<heller>
jbjnr: yes
<heller>
Fix it ;)
<jbjnr>
hastag "not my job"
<jbjnr>
:)
<jbjnr>
tutorial?
<jbjnr>
skyep?
<jbjnr>
skype?
simbergm has quit [Ping timeout: 240 seconds]
<github>
[hpx] biddisco created default_pool_rename (+2 new commits): https://git.io/vdnNT
<github>
hpx/default_pool_rename fe5a5e6 John Biddiscombe: Allow default pool to be renamed...
<github>
hpx/default_pool_rename 1f8770f John Biddiscombe: Add test for pool naming and executor use
<github>
[hpx] biddisco opened pull request #2926: Default pool rename (master...default_pool_rename) https://git.io/vdnNm
<github>
[hpx] biddisco pushed 1 new commit to fixing_2918: https://git.io/vdnpq
<github>
hpx/fixing_2918 ff2e712 John Biddiscombe: Fix Copyright
simbergm has joined #ste||ar
<pree>
what's the difference between hpx::lcos and hpx::lcos::local in most cases ?
<pree>
whether lcos::local means the control objects will work in the current locality ?
simbergm has quit [Ping timeout: 248 seconds]
david_pfander1 has joined #ste||ar
david_pfander1 has quit [Ping timeout: 255 seconds]
simbergm has joined #ste||ar
<heller>
pree: yes. Only local operations exposed
<pree>
heller : whehther local operations exposed means the lightweight control objects only synchronize the tasks which are running local (on current locality ) ?
<pree>
because I'm receiving some errors in local::channel
<pree>
sort of clarification : )
<pree>
thanks
<heller>
Which errors?
<pree>
heller : Sorry got to go, I will tell later today , bye
pree has quit [Quit: AaBbCc]
EverYoung has joined #ste||ar
simbergm has quit [Ping timeout: 246 seconds]
EverYoung has quit [Ping timeout: 255 seconds]
kisaacs has joined #ste||ar
kisaacs has quit [Ping timeout: 248 seconds]
simbergm has joined #ste||ar
hkaiser has joined #ste||ar
jbjnr has quit [Read error: Connection reset by peer]
jbjnr has joined #ste||ar
K-ballo has joined #ste||ar
<github>
[hpx] biddisco pushed 1 new commit to default_pool_rename: https://git.io/vdcnV
<github>
hpx/default_pool_rename a328a77 John Biddiscombe: Fix inspect report
hkaiser has quit [Read error: Connection reset by peer]
aserio has joined #ste||ar
parsa has joined #ste||ar
denis_blank has joined #ste||ar
eschnett has quit [Quit: eschnett]
hkaiser has joined #ste||ar
eschnett has joined #ste||ar
<jbjnr>
circle ci is very annoying and takes far too long
<jbjnr>
heller: project d69 has expired - you are going to be locked out. Hoping to get an extension.
<K-ballo>
yeah, cycle times grew out of control
<jbjnr>
the cycle times are very annoying, but more annoying is that frequently the fails are due to system problems rather than code issues - also, I think we should run inspect first! I'm fed up with waiting 4 hours to find a #include missing and then another 4 to fix it. Yes. I should run inspect locally first. I know.
kisaacs has joined #ste||ar
<K-ballo>
yeah, inspect first would make more sense, but it's entangled with hpx
<K-ballo>
should be possible to move it forward, after core, easily
<K-ballo>
I seem to recall we've tried that already
<jbjnr>
hmmm
<jbjnr>
we should keep a binary around and use it from a prebuilt - it doesn't change very often does it
<jbjnr>
only K-ballo ever updates it!
simbergm has quit [Ping timeout: 258 seconds]
hkaiser has quit [Ping timeout: 246 seconds]
rod_t has joined #ste||ar
kisaacs has quit [Ping timeout: 260 seconds]
Matombo has joined #ste||ar
kisaacs has joined #ste||ar
hkaiser has joined #ste||ar
EverYoung has joined #ste||ar
pree has joined #ste||ar
pree has quit [Remote host closed the connection]
pree has joined #ste||ar
simbergm has joined #ste||ar
diehlpk_work has joined #ste||ar
kisaacs has quit [Ping timeout: 255 seconds]
denis_blank has quit [Ping timeout: 240 seconds]
vamatya has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
kisaacs has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
kisaacs has quit [Ping timeout: 248 seconds]
parsa has joined #ste||ar
pree has joined #ste||ar
pree has quit [Ping timeout: 264 seconds]
simbergm has quit [Ping timeout: 248 seconds]
pree has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
<heller>
I want to replace it with a clang based solution eventually
aserio has quit [Ping timeout: 258 seconds]
<heller>
clang-tidy checks sound perfect for this
wash has joined #ste||ar
pree has quit [Ping timeout: 258 seconds]
kisaacs has joined #ste||ar
kisaacs has quit [Ping timeout: 246 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
pree has joined #ste||ar
<pree>
heller : Hi ,Error regarding the channel is " it fails to compile when using float as template parameter " but works with typedef some_t = float
<heller>
the problem is, that channel also registers the LCO values, which conflicts with the pre registered ones
<heller>
we really need a better documentation there and probably pre register a consistent set of types with those classes
<heller>
a workaround is to give it a different name
<heller>
in your case, HPX_REGISTER_CHANNEL(float_t) will expand to something like HPX_REGISTER_CHANNEL(float_t, float_t). The second parameter is the name
<heller>
which can be chosen arbitrarily, but should be unique across the program
<pree>
heller : Okay, I'm add an issue in github so that it gets (wide) attentions
<pree>
*attention
<pree>
thanks
<hkaiser>
pree: this should have been resolved a while back, what HPX version do you use?
parsa[w] has quit [Read error: Connection reset by peer]
<heller>
it requires a fix. a fix to that problem isn't exactly trivial, i think
<hkaiser>
pree use top of master from the repo, that should fix thi sissue
<heller>
ahh, i see
<heller>
right, this fixes it
<pree>
hkaiser : okay
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
david_pfander has quit [Ping timeout: 246 seconds]
parsa has quit [Client Quit]
parsa has joined #ste||ar
<heller>
will john return tonight?
<hkaiser>
heller: return where?
<heller>
return to IRC ;)
<hkaiser>
shrug - he might
<heller>
from wherever he is right now
mbremer has joined #ste||ar
<heller>
ok, i'll hang around a little longer
jaafar_ has joined #ste||ar
simbergm has joined #ste||ar
Matombo has quit [Ping timeout: 264 seconds]
aserio has joined #ste||ar
Matombo has joined #ste||ar
simbergm has quit [Ping timeout: 240 seconds]
jbjnr_ has joined #ste||ar
<jbjnr_>
hkaiser: I'm here. my desktop machine crashed and I lost my remote dsktop. I can't restart it until I go back to the office tomorrrow
aserio has quit [Ping timeout: 258 seconds]
patg[[w]] has joined #ste||ar
EverYoung has joined #ste||ar
<github>
[hpx] biddisco pushed 1 new commit to default_pool_rename: https://git.io/vdCfY
<github>
hpx/default_pool_rename 91fbdb9 John Biddiscombe: Use get_default_pool_name() internally for consistency
<hkaiser>
jbjnr_: heller was looking for you
<jbjnr_>
that hello world test is behaving oddly here too. I think there is a race in something
simbergm has joined #ste||ar
<jbjnr_>
I don't think it's related to my changes though
aserio has joined #ste||ar
aserio1 has joined #ste||ar
<jbjnr_>
hkaiser: I sometimes get hello world from OS-thread 0 on locality 0 hello world from OS-thread 1 on locality 0 hoello world from OS-thread 576 kB Node 0 MemFree: 0 on locality 1 hoello world from OS-thread 576 kB Node 0 MemFree: 0 on locality 1
<jbjnr_>
There are extra "Memfree 0" outputs. no ide where that is coming from
aserio has quit [Ping timeout: 248 seconds]
aserio1 is now known as aserio
<jbjnr_>
and sometimes more than 2 outputs per locality - using master branch
<heller>
jbjnr_: still around and up for skype?
<heller>
jbjnr_: is this coming from the job launcher?
<jbjnr_>
heller: I don't think so now. But we need to agree on who does what
<jbjnr_>
please edit my google doc with what you will do!
<heller>
jbjnr_: ok, I have some questions regarding the slots
<jbjnr_>
feel free to move them around
hkaiser has quit [Read error: Connection reset by peer]
<heller>
there is session 3, 4 and 4 currently
<heller>
all happening at the same time
hkaiser has joined #ste||ar
<hkaiser>
jbjnr_: yah, something seems to be off
<jbjnr_>
yes. I started editing then got confused and need feedback from you, just edit away!
<heller>
I am happy to do the stencil stuff on day one, and have all the advanced stuff, GPU and RP etc. day two
<jbjnr_>
ok, then sess 3 : options, building and running. me, sess 4, stencil you
<heller>
and then have the stencil as the worked example that they will extend to use the features discussed, what do you think?
<heller>
sounds good
<heller>
RP you, GPU me
<heller>
that's the morning of the friday, I guess
<jbjnr_>
ok
<jbjnr_>
hkaiser: yes. I susoect something fishy with master today
<heller>
and in the afternoon, we show the profiling and stuff
<jbjnr_>
yes,
<heller>
great. that'll work
<heller>
we celebrate our great annection tomorrow
<jbjnr_>
I will do a session on "policies, executors, pools and algorithms" and introduce RP, then do the resource management stuff (Fri morning), and we fit gpus around it
<jbjnr_>
annection?
<heller>
some also call it reunion
<jbjnr_>
ah. east/west rejoined
<heller>
yes
<jbjnr_>
got it
<jbjnr_>
we call it reunification
<heller>
annesation
<heller>
annexation
<heller>
anyways
<heller>
reunification sounds great
<heller>
I am sure I can make the code needed for the example work and adapt the remaining sessions on wednesday
<jbjnr_>
ok,
<heller>
don't panic.
<jbjnr_>
too lat
<heller>
it would be boring otherwise ;)
<jbjnr_>
late
<heller>
sorry...
EverYoung has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
<jbjnr_>
yup. hello world on threads>1 is broken somehow
<github>
hpx/master 15c4a91 Hartmut Kaiser: Fixing tests relying on static_partitioner
<github>
hpx/master b29dc43 Hartmut Kaiser: Fixing usage of some of the executor parameters customization points
<github>
[hpx] aserio pushed 2 new commits to add_checkpoint: https://git.io/vdCq2
<github>
hpx/add_checkpoint 4f93190 aserio: Revisions per pull request comments...
<github>
hpx/add_checkpoint 83ad3c6 aserio: Merge branch 'master' into add_checkpoint
<jbjnr_>
ok I would like to merge the RP fixes from you and the custom pool changes from me
<hkaiser>
jbjnr_: shouldn
<hkaiser>
't we do that one at a time?
<jbjnr_>
I've been using this code pretty heavily over the last ffident it's okew days and I am fairly con
<jbjnr_>
I've been using this code pretty heavily over the last few days and I am fairly confident it's ok
<jbjnr_>
bloody trackpad
<jbjnr_>
but one at a time works for me
<github>
[hpx] biddisco closed pull request #2919: Make sure the resource_partitioner is properly destructed on hpx::finalize (master...fixing_2918) https://git.io/vdZls
<github>
[hpx] biddisco force-pushed fixing_2914 from 2e59f03 to 9696aae: https://git.io/vdnqV
<github>
hpx/fixing_2914 a1c864e John Biddiscombe: Custom pool executor was missing priority and stacksize options
<github>
hpx/fixing_2914 9696aae John Biddiscombe: Rename customized_pool_executor to pool_executor
zbyerly_ has joined #ste||ar
wash has quit [Remote host closed the connection]
kisaacs has joined #ste||ar
kisaacs has quit [Ping timeout: 248 seconds]
EverYoung has joined #ste||ar
<diehlpk_work>
what(): partitioner::add_resource: Creation of 8 threads requested by the resource partitioner, but only 7 provided on the command-line.
<diehlpk_work>
Can anyone explain what this means?
simbergm has quit [Ping timeout: 240 seconds]
<aserio>
jbjnr_: ^^
mbremer has quit [Quit: Page closed]
jaafar_ has joined #ste||ar
<diehlpk_work>
I got this error with hpx master
<diehlpk_work>
With the version of last week on my local machine my code is working
kisaacs has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
<diehlpk_work>
aserio, Should I gnerate a ticket for that?
<diehlpk_work>
jbjnr_, I used a fresh build from current master and still there
kisaacs has quit [Ping timeout: 240 seconds]
<diehlpk_work>
jbjnr_, When using 64 it is running
<jbjnr_>
does it give this error with hello world etc?
<diehlpk_work>
Have to build examples first, will let you know soon
EverYoun_ has joined #ste||ar
<jbjnr_>
(just a make -j hello_world_exe should do it)
vamatya_ has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
<diehlpk_work>
jbjnr_, Yes
<diehlpk_work>
Without -t working
<diehlpk_work>
-t 64 too
<diehlpk_work>
1 to 63 is not working
<diehlpk_work>
Same for fibo
vamatya has quit [Ping timeout: 255 seconds]
<jbjnr_>
shit - looks like hartmut broke master cos my stuff is not building and it's the hpx::parallel::execution::detail::customization_point that heller was working on that fails
<jbjnr_>
hello world built though. let me test
<jbjnr_>
works fine for me
<heller>
what's wrong?
<diehlpk_work>
-t 1 to 63 results into error
EverYoun_ has quit [Remote host closed the connection]
<diehlpk_work>
what(): partitioner::add_resource: Creation of 4 threads requested by the resource partitioner, but only 3 provided on the command-line.
<jbjnr_>
please give me full hello world command line
EverYoung has joined #ste||ar
<jbjnr_>
diehlpk_work: ^
<heller>
diehlpk_work: more than 64 CPUs in your system?
<heller>
diehlpk_work: configure HPX with -DHPX_WITH_MAX_CPUS=256
<jbjnr_>
yes
<diehlpk_work>
bin/hello_world -t 1
<diehlpk_work>
terminate called after throwing an instance of 'std::runtime_error'
<diehlpk_work>
what(): partitioner::add_resource: Creation of 2 threads requested by the resource partitioner, but only 1 provided on the command-line.
<diehlpk_work>
Aborted
<diehlpk_work>
heller, 128
<heller>
diehlpk_work: -DHPX_WITH_MAX_CPUS=128
<jbjnr_>
and HPX_WITH_MORE_THAN_64_THREADS=ON
<heller>
that's not needed
<jbjnr_>
no?
<heller>
nop
<heller>
nope
<jbjnr_>
is is obsolete now
<heller>
guess so
<jbjnr_>
can we get rid of it then
<heller>
yes
<jbjnr_>
good
<diehlpk_work>
Yes, when I start hello_work without -t all 128 are used
<heller>
yes
<jbjnr_>
diehlpk_work: that is very important the -DHPX_WITH_MAX_CPUS=128 if that is not set then everything is wrong in the masks so you will get a ton of errors
<heller>
mostly integer overflows and such
<heller>
jbjnr_: since when is yours broken?
<diehlpk_work>
Ok, I will try if this solves the -t error
EverYoun_ has joined #ste||ar
<jbjnr_>
since hartmut pushed to master an hour ago
<jbjnr_>
we should put him in the stocks
<heller>
yes
<heller>
I have no idea what's going on there
<heller>
he said he wanted to fix something
jaafar_ has quit [Ping timeout: 240 seconds]
<heller>
I didn't even look at the parameters thingy before he merged it
<heller>
so everything is broken now
EverYoung has quit [Ping timeout: 246 seconds]
<heller>
it's late here, let him fix it once he is back and we'll start on fresh tomorrow morning
kisaacs has joined #ste||ar
<jbjnr_>
not everything, but some. late for me too. I'll send him a message and check back tomorrow
<K-ballo>
what's the parameters thingy?
<diehlpk_work>
jbjnr_, Recompiled with -DHPX_WITH_MAX_CPUS=128
<diehlpk_work>
But same error
<heller>
unlikely
<diehlpk_work>
K-ballo, terminate called after throwing an instance of 'std::runtime_error'
<diehlpk_work>
what(): partitioner::add_resource: Creation of 2 threads requested by the resource partitioner, but only 1 provided on the command-line.
<diehlpk_work>
Aborted
<diehlpk_work>
This thing here
<diehlpk_work>
Sorry have to go home.
kisaacs has quit [Ping timeout: 260 seconds]
kisaacs has joined #ste||ar
jbjnr_ has quit [Ping timeout: 255 seconds]
jaafar_ has joined #ste||ar
hkaiser has joined #ste||ar
kisaacs has quit [Ping timeout: 240 seconds]
<heller>
hkaiser: hey
<heller>
hkaiser: looks like your change broke johns stuff