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/ | GSoC2018: https://wp.me/p4pxJf-k1
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
hkaiser has quit [Quit: bye]
K-ballo has quit [Quit: K-ballo]
nanashi55 has quit [Ping timeout: 268 seconds]
nanashi55 has joined #ste||ar
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Client Quit]
hkaiser has joined #ste||ar
david_pfander has joined #ste||ar
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] msimberg opened pull request #3400: Fix floating point exception scheduler_base idle backoff (master...fix-exponential-backoff) https://github.com/STEllAR-GROUP/hpx/pull/3400
ste||ar-github has left #ste||ar [#ste||ar]
daissgr has joined #ste||ar
jgolinowski has joined #ste||ar
<jgolinowski> jbjnr, should I keep 2 space indent in the MartyCam "spin-off"
<jbjnr> jgolinowski: however you like. It's an old project and the style needs updating anyway
<jbjnr> I think I awa a PR from you, I'd better take a look
<jbjnr> ^saw
<jgolinowski> jbjnr, so the PR is deprecated
<jbjnr> ok
<jbjnr> I will play with it soon.
<jgolinowski> too many things required changes so I started a new project and was copying bit by bit
<jbjnr> :(
<jgolinowski> this way it was easier
<jbjnr> :(
<jgolinowski> now the plan is to wrap it up and put back to PR
<jgolinowski> so you can easily see what changed
<jbjnr> ok
ste||ar-github has joined #ste||ar
<ste||ar-github> [hpx] hkaiser pushed 3 new commits to master: https://github.com/STEllAR-GROUP/hpx/compare/389ca4a24854...fdc279a748e1
<ste||ar-github> hpx/master a6c1c02 Nikunj Gupta: Adds Debug option for hpx initializing from main
<ste||ar-github> hpx/master bd77d6e Nikunj Gupta: Changes according to review
<ste||ar-github> hpx/master fdc279a Hartmut Kaiser: Merge pull request #3385 from NK-Nikunj/Debug_hpx_main...
ste||ar-github has left #ste||ar [#ste||ar]
<jbjnr> jgolinowski: got it thanks
nikunj has joined #ste||ar
<nikunj> zao: yt?
<zao> Not really, getting ready to bake some cookies.
<zao> How's it going?
<nikunj> great!
<nikunj> Could you try building #3367 in your free time?
<nikunj> It's the mac osx implementation
<zao> Not sure if I’ll have much “free time”, lots of stuff going on.
<nikunj> zao, ok no worries
<nikunj> hkaiser, where should I put in my section in the docs?
david_pfander1 has joined #ste||ar
<nikunj> I think it should be in between application and names
<nikunj> applications^
david_pfander has quit [Ping timeout: 244 seconds]
david_pfander1 is now known as david_pfander
heller has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
heller has joined #ste||ar
<hkaiser> nikunj: yah
<nikunj> alright, then I'll include it there. I can't build docs rn since I do not meet the software requirements. I think adding am include there should build it
<hkaiser> nikunj: ok, I'll try building it
<nikunj> hkaiser, thanks
<nikunj> hkaiser, commit added
<hkaiser> ok
mcopik has joined #ste||ar
david_pfander1 has joined #ste||ar
david_pfander has quit [Ping timeout: 240 seconds]
david_pfander1 is now known as david_pfander
mcopik has quit [Ping timeout: 240 seconds]
<zao> nikunj: https://i.imgur.com/AhLzozg.jpg <- way better than HPX
<zao> (yes, applied too much jam)
mcopik has joined #ste||ar
<nikunj> zao, looks delicious!
<jbjnr> zao: please send biscuits right away.
nikunj_ has joined #ste||ar
nikunj_ has quit [Client Quit]
nikunj[m] has joined #ste||ar
nikunj[m] has quit [Client Quit]
nikunj[m] has joined #ste||ar
nikunj[m] has quit [Ping timeout: 252 seconds]
daissgr has quit [Quit: WeeChat 1.9.1]
eschnett has joined #ste||ar
hkaiser has quit [Quit: bye]
K-ballo has joined #ste||ar
aserio has joined #ste||ar
<ms[m]1> hkaiser_: it seems like all hpx examples and tests build cleanly with cuda 9, gcc 6 and c++11 and cmake 11... do you think it would be possible to get that setup running on rostam instead of the current cuda setup? as it is we know that hpx doesn't compile with cuda 8 (it seems to be at least be a problem related to boost), but at least we could make sure we don't introduce regressions for cuda 9
<ms[m]1> pycicle on daint would of course do its share of testing as well
<diehlpk_work> ms[m]1, see pm
<zao> jbjnr: CI clearly stands for cookie incentives.
<jbjnr> Cookmore Immediately!
<K-ballo> mmh, cookies
<diehlpk_work> ms[m]1, Had also some boost issues with HPX and CUDA
<zao> nikunj: Spun off a Debug build, gonna take ages.
<diehlpk_work> nikunj, see pm
<ms[m]1> diehlpk_work: yeah, it seems pretty bad... worst one is the bug in cmake that cause nvcc builds to fail
<diehlpk_work> I had to play around with boost, gcc, and cuda version to get HPXCL compiled
david_pfander1 has joined #ste||ar
david_pfander has quit [Ping timeout: 256 seconds]
david_pfander1 is now known as david_pfander
daissgr has joined #ste||ar
aserio has quit [Ping timeout: 255 seconds]
<ms[m]1> aserio: yt?
aserio has joined #ste||ar
david_pfander1 has joined #ste||ar
david_pfander has quit [Ping timeout: 240 seconds]
david_pfander1 is now known as david_pfander
<jgolinowski> ms[m]1, yt?
jgolinowski has quit [Ping timeout: 276 seconds]
aserio has quit [Ping timeout: 240 seconds]
nikunj_ has joined #ste||ar
aserio has joined #ste||ar
<nikunj_> zao, did the results arrive?
jbjnr has quit [Ping timeout: 265 seconds]
<zao> Still hearing fan noise from the machine, it’s building all tests and examples
<zao> I forget which arch we had stuck tests on before?
<nikunj_> Powerpc
jgolinowski has joined #ste||ar
<zao> No, one of mine.
<nikunj_> I don't remember, sorry
david_pfander has quit [Ping timeout: 244 seconds]
aserio1 has joined #ste||ar
aserio has quit [Ping timeout: 256 seconds]
aserio1 is now known as aserio
hkaiser has joined #ste||ar
<zao> [1599/1723] Linking CXX executable bin/thread_yield_test
nikunj_ has quit [Ping timeout: 252 seconds]
<zao> nikunj: Things built at least.
nikunj[m] has joined #ste||ar
<nikunj[m]> zao, was it an error you quoted?
<zao> No, just progress information.
<nikunj[m]> Ohk, will take another hour from the looks of it
<zao> Running test suite now, but no idea if it'll hang arbitrarily :)
jgolinowski has quit [Read error: Connection timed out]
<zao> It begins... stuck on Start 80: tests.regressions.lcos_dir.wait_all_hang_1946
jgolinowski has joined #ste||ar
<nikunj[m]> K
<nikunj[m]> Rerunning it later might fix it
<zao> Not keen on babysitting this gosh-darn test suite.
<zao> Still don't understand why timeouts don't take care of it.
<zao> And there I accidentally ^C the whole thing.
<zao> nikunj[m]: I can't run this fucking thing, it hangs every bunch of tests.
<zao> Either HPX is horribly broken on macOS, HPX'
<zao> 's test suite is horribly broken on macOS, or your stuff is horribly broken.
<zao> I can't tell, and I don't care to find out.
<zao> You're on your own.
nikunj[m] has quit [Ping timeout: 252 seconds]
hkaiser has quit [Ping timeout: 256 seconds]
nikunj[m] has joined #ste||ar
<nikunj[m]> zao, alright
<nikunj[m]> I'll run it on my machine to check if I'm experiencing the same problem
hkaiser has joined #ste||ar
<nikunj[m]> I didn't change much from my last implementation really
<nikunj[m]> Which ran fine on your machine iirc
<nikunj[m]> I'll try to run it tonight
<nikunj[m]> And find if it's related to my implementation
parsa has joined #ste||ar
aserio has quit [Ping timeout: 256 seconds]
nikunj[m] has quit [Ping timeout: 252 seconds]
<heller> hkaiser: looks like the new future atomic state might have a problem
<hkaiser> ok
<heller> see zaos message
<hkaiser> heller: did he run this branch?
<heller> ah right, we didn't merge it yet
<heller> good point
<heller> thought we did ;)
<heller> guess I should eventually start cross compiling the tests and see how far I get ;)
<heller> hkaiser: appear.in in 5 minutes?
<hkaiser> heller: yes
<hkaiser> heller: brb, need coffee
<zao> heller: Had problems on macOS the last time too, which was like a month ago.
<zao> Assorted tests go and twiddle their thumbs forever.
<zao> This time, something like 4-5 of them in the first 150 or so.
<zao> In short, test suite is unusable.
aserio has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
aserio has quit [Ping timeout: 256 seconds]
<heller> hkaiser: read the whole article though
aserio has joined #ste||ar
parsa has joined #ste||ar
daissgr has quit [Ping timeout: 240 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
<nikunj> zao, HPX is building correctly. Running tests rn
<nikunj> will share the results with you
mcopik has quit [Ping timeout: 240 seconds]
parsa has joined #ste||ar
aserio has quit [Ping timeout: 276 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
aserio has joined #ste||ar
<heller> woohooo! HPX is able to fully utilize the bandwidth on my ARM cluster with a broadcast :D
<zao> utilize or waste? :)
<diehlpk_work> heller, What kind of arm boards do you use?
<heller> nano PC T3
jbjnr has joined #ste||ar
<heller> diehlpk_work: ARM A53, 8 cores, connected over 10Mbit ethernet
<heller> sorry, 100 Mbit
<diehlpk_work> How many boards?
<heller> 16
<diehlpk_work> At your home?
<heller> diehlpk_work: https://pasteboard.co/Hy7exZp.jpg
<heller> diehlpk_work: small demonstrator we built at work
<heller> 2.5k of net worth for the hardware. We build you one for an additional 650 ;)
<diehlpk_work> Cool, costs?
<heller> that's 3.15k, including software setup (SLURM and HPX)
<diehlpk_work> Cool
<diehlpk_work> So one of them is a headnode or?
<heller> yes
<heller> it is completely self contained
<diehlpk_work> What is the usage?
<heller> the primary cause we built it is to have a demonstrator for AllScale
<heller> it's essentially my own little 16 node cluster now :P
<heller> once the project is done, I want to use it for events to lure pupils into computer science
<heller> and testing machine
<diehlpk_work> Sounds good
<heller> so I want to revive my van gogh demo on there
<heller> and eventually measure power
<heller> this device might be very helpful for tsunami forecasts etc
<diehlpk_work> Yes
<heller> (that's actually the intention of our first "customer" here)
<diehlpk_work> It sounds very interesting
<diehlpk_work> Do you have a paper?
<diehlpk_work> You should write a blog post for the ste||ar blog
<diehlpk_work> So we would not have only GSoC posts there
<heller> yes, I should
<heller> there are so many things which I should do
<heller> no, there is no paper
<diehlpk_work> Writing the HPX book :)
<heller> sure
<heller> it'll get published by the end of the year
eschnett has quit [Quit: eschnett]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
aserio has quit [Quit: aserio]
hkaiser has quit [Quit: bye]
jgolinowski has left #ste||ar ["Leaving"]
daissgr has joined #ste||ar
jgolinowski has joined #ste||ar
hkaiser has joined #ste||ar
jgolinowski has quit [Quit: Leaving]
daissgr has quit [Quit: WeeChat 1.9.1]
daissgr has joined #ste||ar