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/
EverYoun_ has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
Smasher has quit [Ping timeout: 240 seconds]
daissgr has quit [Quit: WeeChat 1.4]
Smasher has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 255 seconds]
cogle has joined #ste||ar
EverYoung has joined #ste||ar
hkaiser has quit [Quit: bye]
EverYoung has quit [Ping timeout: 265 seconds]
cogle has quit [Quit: Page closed]
cogle has joined #ste||ar
parsa has joined #ste||ar
Shikhar has joined #ste||ar
Shikhar has quit [Client Quit]
eschnett has joined #ste||ar
hkaiser has joined #ste||ar
vamatya has quit [Ping timeout: 265 seconds]
bibek has quit [Quit: Leaving]
hkaiser has quit [Quit: bye]
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]
quaz0r has quit [Ping timeout: 255 seconds]
quaz0r has joined #ste||ar
parsa has joined #ste||ar
vamatya has joined #ste||ar
mcopik has quit [Ping timeout: 268 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
Anushi1998 has joined #ste||ar
vamatya has quit [Ping timeout: 248 seconds]
cogle has quit [Ping timeout: 260 seconds]
Smasher has quit [Ping timeout: 240 seconds]
Smasher has joined #ste||ar
Anushi1998 has quit [Ping timeout: 252 seconds]
jaafar has quit [Ping timeout: 240 seconds]
Smasher has quit [Ping timeout: 240 seconds]
Smasher has joined #ste||ar
david_pfander has joined #ste||ar
CaptainRubik has joined #ste||ar
<CaptainRubik>
Hi
<CaptainRubik>
Trying to install hpx
<CaptainRubik>
Running Cmake gives the following error
<CaptainRubik>
WARNING: Target "parcel_coalescing_lib" requests linking to directory "/Users/Shikhar/Desktop/hwloc_bin/lib". Targets may link only to libraries. CMake is dropping the item.
marco has joined #ste||ar
<CaptainRubik>
running on high sierra
<simbergm>
hi CaptainRubik, sounds like cmake is expecting the full path to libhwloc and has a directory instead
<simbergm>
how do you specify the hwloc path to cmake?
<CaptainRubik>
Thanks. I figured that out.
jbjnr has joined #ste||ar
<simbergm>
I usually do the equivalent of -DHWLOC_ROOT="/Users/Shikhar/Desktop/hwloc_bin/"
<CaptainRubik>
When i give cmake command is it normal for some tests to fail
<CaptainRubik>
-- Performing Test HPX_WITH_CXX_FLAG_WERROR_TRAMPOLINES - Failed -- "-Werror=trampolines" not available for language CXX.
<CaptainRubik>
^For ex
<simbergm>
these can fail depending on your compiler and version, hpx do things differently if some features are not available
<CaptainRubik>
Alright Thanks.
<CaptainRubik>
on doing make I am receiving the following warnings
<CaptainRubik>
clang: warning: argument unused during compilation: '-pthread' [-Wunused-command-line-argument]
<jbjnr>
CaptainRubik: just a heads up - I am ,listed as one of the mentors for the concurrent data structures project - but this week is school holiday and I'm not at work, hopefully some of the others on here will be around to answer questions specific to that project.
<jbjnr>
NB. cmake/compile warnings - don't wory too much - errors - they matter
<CaptainRubik>
Okay. (y)
<jbjnr>
my advice is do a make -j8 hello_world_exe first - check that that binary runsa and works - then do a make tests or something - otherwise the build/test/rebuild with new flags cycle can take way too long
<jbjnr>
I think we need a wiki entry for new users to get going with building/running/testing.
Smasher has quit [Quit: Connection reset by beer]
Smasher has joined #ste||ar
<jbjnr>
just fyi - if you have problems with builds - have a peek at https://github.com/biddisco/biddisco.github.io/wiki/Linux (laptop debug build) and there are other pages on the same wiki with my settings for other machines that I build hpx on.
<CaptainRubik>
Will try to make hello_world_exe
<CaptainRubik>
Thanks for the resources
<jbjnr>
tip : make help - this will show you all targets
<zao>
Good morning.
<jbjnr>
g'morning
<CaptainRubik>
Good Morning
<jbjnr>
I'm about to go skiing soon, so will disappear in a bit
<CaptainRubik>
Enjoy :)
<simbergm>
jbjnr: do you have time to check out pycicle?
<jbjnr>
what's up?
<jbjnr>
broken?
<simbergm>
no results
<jbjnr>
hmmmm
<jbjnr>
I'll have a look
<simbergm>
test statuses seem to have been sent to github but nothing on cdash
<simbergm>
thanks
CaptainRubik has left #ste||ar [#ste||ar]
<simbergm>
let me know if you have to go and I'll spin up the old version (without sanitizers)
<jbjnr>
dashboard upgrade not as smooth as expected after all
<jbjnr>
try it now anyway, just to see if I did something silly at my end
<simbergm>
sure
<simbergm>
it is getting results from rostam...
<jbjnr>
looks like the custom tracks are not displaying
<jbjnr>
master/pull requests/inspect
<jbjnr>
ok. it looks as though the builds are running correctly, and submissions are being made to the dashboard, but they are just not being displayed
<jbjnr>
might be a cdash 'bug' in the new version
<jbjnr>
will have to check it out later today.
<simbergm>
okay, thanks for checking it
Shikhar has joined #ste||ar
Shikhar has quit [Client Quit]
<simbergm>
K-ballo: yt? you're the hpx::util::format_to expert, no?
<simbergm>
heller_: what would like done for the throttle example? do I understand it correctly that one can send commands to say "suspend 3", but it doesn't actually suspend now?
<simbergm>
and throttle/hierarchical schedulers, okay to remove now?
<heller_>
simbergm: we are going to tackle the suspension problem on firday
<heller_>
friday
sam29 has joined #ste||ar
Shikhar has joined #ste||ar
Shikhar has quit [Client Quit]
sam29 has quit [Ping timeout: 260 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
daissgr has joined #ste||ar
<simbergm>
heller_: okay, but "tackle the suspension problem" means update the throttle example?
<heller_>
simbergm: err
<simbergm>
I'm confused
<heller_>
me too ;)
<heller_>
updating the throttle example: no
<simbergm>
so both yes means:
<heller_>
I have no idea what it does
<heller_>
yes
<simbergm>
1. schedulers can be removed
<heller_>
I was confused
<simbergm>
2. ?
<heller_>
1. schedulers can be removed
<heller_>
2. I am not familiar enough with the throttle example
<heller_>
BUT, suspending/resuming cores currently doesn't work reliably
<simbergm>
aha, I see
<heller_>
that's what I meant with tackling the suspension problem ... sorry
<simbergm>
so you have some problems with the allscale stuff?
<heller_>
I am jumping between various different contexts at the moment
<heller_>
yes
<simbergm>
ok
<simbergm>
what happens?
<heller_>
it looks like there is a race between actually suspending and resuming
<heller_>
so it appears, that the cores are never actually getting suspended
<simbergm>
it never works or occasionally works?
<heller_>
it seems to never work in our context
<heller_>
the test appears to be ok though
<simbergm>
hrm, that's bad
<heller_>
yeah
<heller_>
We have a debug session planned for friday
<simbergm>
if you have any small examples where it doesn't work I can take a look
<simbergm>
okay
<simbergm>
let me know if there's anything I can do to help
<heller_>
I'll have to get the full picture for myself as well...
<simbergm>
sure
<simbergm>
one thing I've realized is that I haven't protected the suspension from spurious wakeups, but I have no idea how often those actually happen in real life
<heller_>
spurious wake ups as in?
<simbergm>
as in apparently a thing that condition variables can do where they decide to wake up by themselves...
<simbergm>
but it's probably not the reason you're having problems
<simbergm>
just using hpx::cout << "hello ..." works correctly
<K-ballo>
no, they should be identical
<K-ballo>
I imagine hpx::cout only does its magic if used as an hpx::cout, and does not work properly when used as a stream?
<simbergm>
yeah, that sounds like a good guess, I noticed that it never goes to the hpx ostream << overloads when using format_to, but this is then expected behaviour?
<simbergm>
or not surprising at least?
cogle has joined #ste||ar
cogle has quit [Client Quit]
parsa| has quit [Quit: *yawn*]
<K-ballo>
oh, that's surprising.. sounds like incredibly bad design
<K-ballo>
there must be a reason for that
<K-ballo>
but yes, it would explain why it doesn't behave as an hpx stream, since it is being used as just a stream
akheir has joined #ste||ar
tianyi has joined #ste||ar
<K-ballo>
simbergm: see if you can add a format_to overloads for hpx streams (yuck), and if that fails change the format_to calls with hpx streams to plain format calls
<simbergm>
K-ballo: thanks for the pointers, I'll try that out
<simbergm>
with plain format calls you mean hpx::cout << hpx::util::format(...) << ...?
<K-ballo>
yep
<simbergm>
okay, thanks
hkaiser has joined #ste||ar
zombieleet has joined #ste||ar
rtohid has joined #ste||ar
Ashish_ has joined #ste||ar
<Ashish_>
Hello everyone, I am Ashish Jha, junior year undergrad at NIT Rourkela. I am interested in the GSoC project titled HPX Backend for OpenCV. Someone please guide me where to start?
<Ashish_>
I have worked in OpenMP and STL extensively and is beginner in OpenCV
<hkaiser>
hey Ashish_
aserio has joined #ste||ar
<Ashish_>
Hi @hkaiser
<Ashish_>
Is there any other communication medium for Ste||ar?
<simbergm>
hi Ashish_, this is the main place for discussion but there is also the mailing list hpx-users where you can post questions
<simbergm>
I'm not too familiar with the opencv project but a good place to start is always just to get familiar with hpx, building it, running examples and looking around the code
<simbergm>
for specifics on that project I know that at least jbjnr is a good point of contact but I believe he's mostly away this week (he still pops in every now and then)
<Ashish_>
Okay thanks for the inforamtion, simbergm!
zombieleet has quit [Ping timeout: 265 seconds]
zombieleet has joined #ste||ar
akheir has quit [Remote host closed the connection]
hkaiser has quit [Quit: bye]
sam29 has joined #ste||ar
Ashish_ has quit [Ping timeout: 260 seconds]
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar
akheir has joined #ste||ar
<wash[m]>
aserio: I am around today
<aserio>
wash[m]: lol, we have canceled the meeting today due to Ash Wednesday
<wash[m]>
Ah, lol. Next week!
sam29 has quit [Ping timeout: 260 seconds]
david_pfander has quit [Ping timeout: 268 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
zombieleet has quit [Ping timeout: 264 seconds]
zombieleet has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
Smasher has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
Smasher has joined #ste||ar
eschnett has quit [Quit: eschnett]
zombieleet has quit [Ping timeout: 240 seconds]
akheir has quit [Remote host closed the connection]
victor_ludorum has joined #ste||ar
jaafar has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
mcopik has joined #ste||ar
Smasher has quit [Read error: Connection reset by peer]
Smasher has joined #ste||ar
eschnett has joined #ste||ar
<jbjnr>
I do not like the new cdash. it's broken :(
Smasher has quit [Remote host closed the connection]
Smasher has joined #ste||ar
EverYoung has quit [Ping timeout: 260 seconds]
EverYoun_ has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
Smasherr has joined #ste||ar
EverYoung has quit [Ping timeout: 265 seconds]
EverYoung has joined #ste||ar
cogle has joined #ste||ar
sam29 has joined #ste||ar
<sam29>
I'm trying to build hpx, running the cmake command gives the following error "Could NOT find TCMalloc (missing: TCMALLOC_LIBRARY TCMALLOC_INCLUDE_DIR)"
<hkaiser>
sam29: either install tcmalloc or use -DHPX_WITH_MALLOC=system
<sam29>
hkaiser: Thanks, I'll try again with the changes suggested.
aserio1 has joined #ste||ar
simbergm has quit [Ping timeout: 256 seconds]
aserio has quit [Ping timeout: 276 seconds]
aserio1 is now known as aserio
simbergm has joined #ste||ar
victor_ludorum has quit [Quit: Page closed]
<anushi>
Is there any minimum RAM requirement for running make -j5?
sam29 has quit [Ping timeout: 260 seconds]
<anushi>
When I first tried this my system hanged and I have to switch it off and in second try it terminated after half running by saying Virtual memory exhausted
sam29 has joined #ste||ar
<anushi>
sam29: If you are using ubuntu you can try sudo apt-get install libgoogle-perftools-dev
<anushi>
sam29, Also you can look into issue #2524
<zao>
anushi: Hehe... Some tests/examples can eat up to 7-8 GiB each when linking.
<zao>
Otherwise, 1-1.5 GiB per source file, maybe.
<anushi>
zao: I wonder if I will edit some file and build again then I will need to repeat this tiring process again?
<zao>
anushi: As long as you keep your build directory and haven't messed with the CMake files itself, you can just build again in the directory and it'll reuse what it can.
<zao>
If you've modified HPX itself, it'll build the changes, relink the library and relink the tests.
<zao>
If you've modified individual tests or examples, just those will have to build.
<anushi>
Okay, that's great :)
<sam29>
I tried along with the changes, now I'm having following problem "failed to create symbolic link '/home/sam29/packages/hpxIns/my_hpx_build/share/hpx': No such file or directory"
<zao>
That's normal.
<zao>
I believe it's harmless.
<sam29>
zao: Great !! I'll continue further. Thanks !!
<sam29>
heller_: Currently which data structure are we using for lookups of GIDs in associative container. Can you point me to the file where I can find its use in our code?
<heller_>
sam29: it's in multiple places
<heller_>
sam29: addressing_service.hpp and the cpp file
<heller_>
sam29: and primary_namespace
<heller_>
sam29: we currently use std::unordered_map
<sam29>
heller_: Thank you. I'll have a look at the code.
hkaiser has joined #ste||ar
<heller_>
sam29: the goal of the project is to figure out if we could exploit the structure of the GIDs, to get a faster lookup