hkaiser 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/
diehlpk has joined #ste||ar
diehlpk has quit [Ping timeout: 260 seconds]
anushi has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
hkaiser has quit [Quit: bye]
parsa has joined #ste||ar
anushi has quit [Remote host closed the connection]
ravibitsgoa has joined #ste||ar
<ravibitsgoa> Hi everyone,
<ravibitsgoa> I would like to start contributing to hpx library. I know C++ STL and data structures and algorithms. I also know some basics of distributed computing.
<ravibitsgoa> Where should I start from?
parsa has quit [Quit: Zzzzzzzzzzzz]
nanashi55 has quit [Ping timeout: 272 seconds]
nanashi55 has joined #ste||ar
parsa has joined #ste||ar
parsa has quit [Client Quit]
ravibitsgoa has quit [Ping timeout: 256 seconds]
parsa has joined #ste||ar
ravibitsgoa has joined #ste||ar
ravibitsgoa has quit [Ping timeout: 256 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
jbjnr__ has joined #ste||ar
jbjnr_ has quit [Ping timeout: 264 seconds]
jaafar has joined #ste||ar
<jbjnr__> is anyone online?
david_pfander has joined #ste||ar
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco force-pushed template_typename from a4796cc to b4fc860: https://github.com/STEllAR-GROUP/hpx/commits/template_typename
<ste||ar-github> hpx/template_typename ad94ec0 John Biddiscombe: template <typename> typename is c++14 only, use template<typename> class
<ste||ar-github> hpx/template_typename b4fc860 John Biddiscombe: Do not build guided pool examples with older GCC compilers
ste||ar-github has left #ste||ar [#ste||ar]
nikunj has joined #ste||ar
jaafar has quit [Ping timeout: 252 seconds]
anushi has joined #ste||ar
heller_ has quit [Read error: Connection reset by peer]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco force-pushed template_typename from b4fc860 to 3ec11fe: https://github.com/STEllAR-GROUP/hpx/commits/template_typename
<ste||ar-github> hpx/template_typename 3ec11fe John Biddiscombe: Do not build guided pool examples with older GCC compilers
ste||ar-github has left #ste||ar [#ste||ar]
heller_ has joined #ste||ar
<heller_> ha, SASL auth 4tw
anushi has quit [Ping timeout: 252 seconds]
Anushi1998 has joined #ste||ar
Anushi1998 has quit [Client Quit]
Anushi1998 has joined #ste||ar
<jbjnr__> I'm troubled by buildbot. When you force a build, it seems as though it isn't actually building your branch
<jbjnr__> either that, or it is still building ones from a long way back in the queue
Anushi1998 has quit [Ping timeout: 244 seconds]
<heller_> jbjnr__: hmm
Anushi1998 has joined #ste||ar
Anushi1998 has quit [Ping timeout: 268 seconds]
Anushi1998 has joined #ste||ar
<jbjnr__> heller_: for example, I triggered build this morning and get this http://rostam.cct.lsu.edu/builders/hpx_gcc_4_9_boost_1_58_centos_x86_64_release
<jbjnr__> and the sha it is building, is not my branch
<jbjnr__> :(
<jbjnr__> ooh.
<jbjnr__> The SingleBranchScheduler scheduler named 'hpx_continuous_integration' triggered this build
<jbjnr__> so it was not the one I triggered. Then I must wait longer.
<heller_> yeah
<jbjnr__> balls. I'm looking at the wrong builder
<zao> :)
<jbjnr__> mine passed \o/ yay
<heller_> the buildbot needs a serious upgrade... still running 0.8x while 1.x is out...
<jbjnr__> <cough> pycicle </cough>
<zao> jbjnr__: Are the two command lines supposed to be identical in the readme? https://github.com/biddisco/pycicle#running-pycicle
<jbjnr__> doh!
<jbjnr__> no the second one should say dca instead of hpx
<heller_> jbjnr__: yes!
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco force-pushed template_typename from 3ec11fe to c58d32c: https://github.com/STEllAR-GROUP/hpx/commits/template_typename
<ste||ar-github> hpx/template_typename c58d32c John Biddiscombe: Do not build guided pool examples with older GCC compilers
ste||ar-github has left #ste||ar [#ste||ar]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco deleted template_typename at c58d32c: https://github.com/STEllAR-GROUP/hpx/commit/c58d32c
ste||ar-github has left #ste||ar [#ste||ar]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco created guided_pool_gcc_5 (+1 new commit): https://github.com/STEllAR-GROUP/hpx/commit/c58d32c8a03f
<ste||ar-github> hpx/guided_pool_gcc_5 c58d32c John Biddiscombe: Do not build guided pool examples with older GCC compilers
ste||ar-github has left #ste||ar [#ste||ar]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] biddisco opened pull request #3524: Do not build guided pool examples with older GCC compilers (master...guided_pool_gcc_5) https://github.com/STEllAR-GROUP/hpx/pull/3524
ste||ar-github has left #ste||ar [#ste||ar]
<zao> jbjnr__: I should totally take a look at pycicle and steal the repository polling parts :)
<jbjnr__> it's very simple. there's a pygithub module that does all the work
<zao> (I'm still noodling around with my test thingie)
hkaiser has joined #ste||ar
ravibitsgoa has joined #ste||ar
<ravibitsgoa> hey, is anyone online to help out a beginner?
<hkaiser> hey ravibitsgoa
<ravibitsgoa> hi, @hkaiser, How should I get started in contributing to hpx?
<ravibitsgoa> I know C++ STL and distributed computing.
<hkaiser> ravibitsgoa: read the docs https://stellar-group.github.io/hpx/docs/sphinx/branches/master/html/index.html, looks through open tickets to see what's interesting to you
<ravibitsgoa> cool thanks.
<hkaiser> we have also a list of potential projects here: https://github.com/STEllAR-GROUP/hpx/wiki/GSoC-2018-Project-Ideas
<hkaiser> some of them are obsolete, but it's a starting point for discussions
<hkaiser> ravibitsgoa: bets is to use a project of your own and try to use hpx there
<hkaiser> best*
<ravibitsgoa> I saw this list. It looks quite intimidating
<hkaiser> heh - no worries
<hkaiser> ravibitsgoa: is there anything in particular interesting to you?
<ravibitsgoa> yes, Adding Lustre Backend to hpxio, Dynamic Load Balancing Using HPX Migration for LibGeoDecomp and STORM Project, and Coroutine-like Interface seem interesteing.
<ravibitsgoa> Are any of these assigned to anyone / done?
<hkaiser> I meant what's particularily interesting for you in general?
<ravibitsgoa> I didn't get you
<hkaiser> what's you general interests?
<hkaiser> algorithms? data structures? parallel control flow?
<hkaiser> API design?
<hkaiser> writing docs?
<ravibitsgoa> data structures and algorithms, distributed systems. not parallel control flow all that much, not API design, neither docs
<hkaiser> ;-)
<hkaiser> figured as much
<ravibitsgoa> I am also a sports programmer BTW
<hkaiser> HPX has a (almost) full implementation of the C++17 parallel algorithms
<hkaiser> here is the full story: https://github.com/STEllAR-GROUP/hpx/issues/1141
<hkaiser> as you can see some are still missing *hint* *hint*
<ravibitsgoa> It seems that other than sorting algorithms, everything is done.
<ravibitsgoa> Okay, I will look into this.
<hkaiser> ravibitsgoa: I also think the list misses a couple of algorithms that have been added later (closer to the C++17 standard - timewise)
<hkaiser> feel free to report on this ticket if you find any missing ones
<hkaiser> ravibitsgoa: there is also the Ranges TS which we have started to adapt towards (see: https://github.com/STEllAR-GROUP/hpx/issues/1668)
<hkaiser> also, there is a newer TS we would like to conform with: https://github.com/STEllAR-GROUP/hpx/issues/3364
<hkaiser> plenty of options...
<ravibitsgoa> What is a TS, BTW? ?
<hkaiser> TS == technical specification - in laymen's terms a 'not-quite-standard-with-standards-committee-approval' ;-)
<ravibitsgoa> Also, where can I find some help about using IRC?
<ravibitsgoa> I am currently facing some trouble
ravibitsgoa has quit [Ping timeout: 256 seconds]
eschnett_ has quit [Quit: eschnett_]
<hkaiser> zao: yah, K-ballo just removed those from HPX
<zao> Heh, nice.
<hkaiser> uhh
<simbergm> not quite sure how the circleci builder succeeds when all pycicle builders fail, so don't know if those errors make sense
<simbergm> ignore the cuda one obviously
<hkaiser> yah, that's what I'm wondering as well
<hkaiser> simbergm: I think I know where to look, give me a day or so, pls
<simbergm> sure, no problem
<simbergm> and if you think it's bogus let me know and I'll look into it on the pycicle side
<hkaiser> there might be a reason why for_loop was implemented the way it is ;-)
<simbergm> maybe...
<simbergm> I wouldn't know though :P
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] msimberg pushed 4 new commits to master: https://github.com/STEllAR-GROUP/hpx/compare/e30a1acec47e...96d792cda422
<ste||ar-github> hpx/master fe1c736 Mikael Simberg: Remove docs symlink and versioned docs folders
<ste||ar-github> hpx/master 330779b Mikael Simberg: Don't install source files into share directory
<ste||ar-github> hpx/master 36c2d29 Mikael Simberg: Add docs to all target
ste||ar-github has left #ste||ar [#ste||ar]
<hkaiser> simbergm: so you made it back safely?
<simbergm> yep! waiting to collapse after sitting next to two small children who refused to sleep on the plane, but otherwise everything went smoothly
<hkaiser> :D
<simbergm> the branch does fail in the same way on circleci btw, I just assumed it was the same old failures with the parallel algorithms: https://circleci.com/gh/STEllAR-GROUP/hpx/41831?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link
<hkaiser> doh! sorry - I should have looked more closely
<simbergm> np :)
<hkaiser> I'll fix it today
hkaiser has quit [Quit: bye]
<simbergm> hkaiser: thanks
<simbergm> do you happen to know why here: https://github.com/STEllAR-GROUP/hpx/blob/96d792cda422766a948d5512961076e07b6d9072/hpx/runtime/actions/lambda_to_action.hpp#L81-L92 the warning was ignored rather than the const added? I'm about to add the const (it's a warning also on linux), but will I break windows builds if I do so?
<simbergm> hkaiser ^
<zao> simbergm: He quit.
<simbergm> oh damn
<simbergm> thanks zao
<simbergm> zao: that would be #3505
<zao> Ah.
<simbergm> which the first time around broke on rostam and now magically seems to work...
eschnett_ has joined #ste||ar
aserio has joined #ste||ar
<simbergm> there's a few error codes and configuration options still around for that
Anushi1998 has quit [Ping timeout: 250 seconds]
Anushi1998 has joined #ste||ar
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://github.com/STEllAR-GROUP/hpx/commit/3b324c85cab2ad6e6e45f4331b9019d26c0d5256
<ste||ar-github> hpx/gh-pages 3b324c8 StellarBot: Updating Sphinx docs
ste||ar-github has left #ste||ar [#ste||ar]
<zao> simbergm: Are these problems with 3521 known? Didn't seem to show in the cdash due to "maximum number of errors"? https://gist.github.com/zao/a65fe24adbf7d72f38250cd4648cbb1e
<zao> I'll see if I can get a full log out :)
hkaiser has joined #ste||ar
<simbergm> zao: looks new, but there's plenty of build errors on that one
<zao> Hehe.
<simbergm> (other build errors)
<simbergm> hkaiser (IRC): do you happen to know why here: https://github.com/STEllAR-GROUP/hpx/blob/96d792cda422766a948d5512961076e07b6d9072/hpx/runtime/actions/lambda_to_action.hpp#L81-L92 the warning was ignored rather than the const added? I'm about to add the const (it's a warning also on linux), but will I break windows builds if I do so?
<K-ballo> simbergm: looks like it, yeah
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] hkaiser force-pushed fixing_parallel_for_loop from 3f5f886 to 35f6476: https://github.com/STEllAR-GROUP/hpx/commits/fixing_parallel_for_loop
<ste||ar-github> hpx/fixing_parallel_for_loop 35f6476 Hartmut Kaiser: Fixing argument handling in induction and reduction clauses for parallel::for_loop
ste||ar-github has left #ste||ar [#ste||ar]
<simbergm> K-ballo: which question are you answering? :)
<K-ballo> the sodium security one
<simbergm> ok, thanks
Anushi1998 has quit [Read error: Connection reset by peer]
<hkaiser> simbergm: don't remember :/
<simbergm> hkaiser (IRC): hmm, ok
<hkaiser> simbergm: but yah, adding const should be fine
jbjnr_ has joined #ste||ar
jbjnr__ has quit [Ping timeout: 246 seconds]
david_pfander has quit [Ping timeout: 240 seconds]
<simbergm> let's give it a try...
<hkaiser> simbergm: I'm working on re-applying the patches from #2859
<simbergm> hkaiser (IRC): very nice, thanks!
Anushi1998 has joined #ste||ar
akheir has joined #ste||ar
<akheir> heller_: yt?
aserio1 has joined #ste||ar
aserio has quit [Ping timeout: 250 seconds]
aserio1 is now known as aserio
<zao> simbergm: GCC 7.3.0 build with C++17 decided to get somewhere... https://gist.github.com/zao/d142fedaaa57adf0010a5310719c50a8
aserio has quit [Ping timeout: 252 seconds]
jaafar has joined #ste||ar
Anushi1998 has quit [Ping timeout: 246 seconds]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] hkaiser created segmented_algorithms (+10 new commits): https://github.com/STEllAR-GROUP/hpx/compare/c27f510b6a32^...0877e695e1d4
<ste||ar-github> hpx/segmented_algorithms c27f510 Ajai V George: Binary Transform Reduce added
<ste||ar-github> hpx/segmented_algorithms f141c5d Hartmut Kaiser: Re-applying https://github.com/STEllAR-GROUP/hpx/commit/2f263bfdd295b465df9af1fe2cb31b652de063fd.patch
<ste||ar-github> hpx/segmented_algorithms 34edc7b Hartmut Kaiser: Re-applying https://github.com/STEllAR-GROUP/hpx/commit/e8be3ae898ed71127e93b455fcab73dc1d596c44.patch
ste||ar-github has left #ste||ar [#ste||ar]
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] hkaiser opened pull request #3525: Segmented algorithms (master...segmented_algorithms) https://github.com/STEllAR-GROUP/hpx/pull/3525
ste||ar-github has left #ste||ar [#ste||ar]
diehlpk_work has joined #ste||ar
<diehlpk_work> heller_, Do you have some instructions for building hpx on arm?
hkaiser has quit [Quit: bye]
aserio has joined #ste||ar
<K-ballo> aserio: ping
<aserio> K-ballo: pong
<aserio> How is the meeting going?
<K-ballo> aserio: see pm
<K-ballo> it's not crazy yet
nikunj has quit [Remote host closed the connection]
hkaiser has joined #ste||ar
eschnett_ has quit [Quit: eschnett_]
aserio has quit [Ping timeout: 268 seconds]
aserio has joined #ste||ar
<aserio> hkaiser: yt?
<hkaiser> here
<aserio> please see pm
aserio has quit [Quit: aserio]
<heller_> akheir: hey
<heller_> akheir: what's up
<zao> Ooh, https://github.com/ehazlett/stellar/ . This would be aptly named if I can run builds in such containers :)
<hkaiser> heller_: akheir needs to talk to you wrt the SC demo
<hkaiser> heller_: will you be able to install it this week (before SC)?
<akheir> heller_: I wanted to ask you what you need on the PC for the SC demo?
<akheir> heller_: first thing is that what you need for OS and what libraries you want?
<heller_> akheir: the usual HPX ones are enough
<heller_> akheir: being able to use cpufreq from the account would be nice too
<heller_> installing and testing the demo on thursday should be no problem
<akheir> by cpufreq you to have the permission?
<heller_> yes
<akheir> I want to wipe a pc and install a new os and give you the root password. how do you like that?
<heller_> ah right, having cpufreq headers and libraries would be good (those are hard to install from userspace)
<heller_> sounds perfect!
<heller_> when would be most convenient for you to checkout and testrun everything?
<akheir> I assume you would be happy with debian? am I right?
<heller_> I don't care
<heller_> but yes, debian would be sufficient
<heller_> and nice
<akheir> If it doesn't matter I like to install the latest fedora to see how it looks.
<heller_> alright
<heller_> sounds good to me
<akheir> but no pressure, your choice
<heller_> I can use fedora just fine :P
<akheir> I will fly on Friday and conveniently scheduled my master defense on Thursday. So I would be happier if we check everything earlier
<akheir> how about Wednesday
<heller_> like wednesday?
<heller_> ok, should be doable
<akheir> ^ cool
<heller_> having the machine available ASAP, would be nice
<heller_> what kind of CPU is in there?
<heller_> and how many RAM?
<akheir> sure. I will provide it by tomorrow
<heller_> perfect
<akheir> two intel haswell E5-2630 v3, each 8 cores/16 threads.
<akheir> 64G RAM
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] hkaiser pushed 1 new commit to segmented_algorithms: https://github.com/STEllAR-GROUP/hpx/commit/202e5c1b3409afbe83375dc6e29ab8b1481edc46
<ste||ar-github> hpx/segmented_algorithms 202e5c1 Hartmut Kaiser: Fixing errors introduced during re-applying the patches from ticket 2859
ste||ar-github has left #ste||ar [#ste||ar]
<akheir> but the graphic card is crappy.
<akheir> do you need a good graphic card?
<heller_> no
<heller_> ok, so I have 16 cores. that's good enough
<akheir> good
<akheir> talk to you tomorrow then
<heller_> yup
<heller_> akheir: do you need my ssh public key?
<heller_> or do you still have it?
<hkaiser> heller_: the new clang 7 container is causing random hpx failure :/
<heller_> ok
<heller_> random as in?
<heller_> can you point me to an example?
<hkaiser> the first build failure is genuine, the rest is random (very worrying) stuff
<akheir> heller_: I have it on Rostam and I will give you a password. I think you should first login to rostam and then login to the new machine. unless if you have a lsu vpn.
<heller_> akheir: ok
<heller_> hkaiser: looking into it
<zao> hkaiser: Smells like it might be OOM:ing?
<hkaiser> very strange assertions
<hkaiser> hmmm
<hkaiser> stack overflows?
<zao> (I'm talking about the failed builds)
<hkaiser> are those disabled on circleci?
<hkaiser> well, I'm not talking about the failed builds
akheir has quit [Quit: Leaving]
<hkaiser> the failed runs are worrying
<heller_> yes
<heller_> I can see those
<heller_> this one for example, right?
<hkaiser> yes, and others
<zao> tests.examples? Since when do we have those? :)
<heller_> are they all of the same form?
<heller_> with the "types not compatible" thing?
<hkaiser> no, all different
<hkaiser> zao: we run the examples as tests, yes
<heller_> what's interesting is that only examples are affected
<hkaiser> heller_: indeed
<hkaiser> smells like stack overflows
<heller_> or some ODR violation
<heller_> which only triggers under some circumstances...
<heller_> let's see
<heller_> hkaiser: I'll look into it tomorrow!
<hkaiser> k
<heller_> have to fix something else first...