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/ | GSoD: https://developers.google.com/season-of-docs/
nikunj has joined #ste||ar
diehlpk has joined #ste||ar
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 246 seconds]
K-ballo1 is now known as K-ballo
eschnett has joined #ste||ar
<hkaiser> parsa: yt?
hkaiser has quit [Quit: bye]
diehlpk has quit [Ping timeout: 272 seconds]
nikunj has quit [Remote host closed the connection]
jaafar has quit [Quit: Konversation terminated!]
jaafar has joined #ste||ar
jaafar has quit [Client Quit]
jaafar has joined #ste||ar
jaafar has quit [Client Quit]
jaafar has joined #ste||ar
jaafar has quit [Client Quit]
jaafar has joined #ste||ar
jbjnr has joined #ste||ar
daissgr has joined #ste||ar
rori has joined #ste||ar
<jbjnr> anyone home?
<K-ballo> welcome back
<jbjnr> hi
<K-ballo> greetings from switzerland
<jbjnr> Hope you're having a nice summer too
<jbjnr> ooh - you're here?
<K-ballo> boudry at the moment
<jbjnr> has the c++ meeting finished
<K-ballo> yes
<jbjnr> any news relevant to hpx?
<jbjnr> quick qeustion ; are we dropping c++11 for HPX
<K-ballo> only relevant bits that come to mind are the sync library and the stop token
<K-ballo> there was also a lot of discussion on executors, futures, etc... as has been usual the last few years
<jbjnr> ok. thanks
<jbjnr> I'd better look up sync library
<jbjnr> I see issue created by mikael for c++14 support, so that's good.
<K-ballo> the sync library is just latches and barriers (which I think we already have) and some atomic stuff
<K-ballo> oh and semaphores, we have those too
<jbjnr> k
<jbjnr> What's hapening in Boudry - are you on vacation, or is this work?
<K-ballo> visiting friends
<jbjnr> great
<simbergm> jbjnr: welcome back!
<jbjnr> hi. Are you at work or on hols too?
<simbergm> there hasn't been a definite decision to require c++14 but I think we can do it for this release, I just haven't done anything about it yet
<simbergm> work
<jbjnr> I'm going to assume that we make c++14 the min anyway then
<jbjnr> it's like a ghost town in the LZA office
<jbjnr> only about 3 of us here.
<simbergm> holiday tomorrow...
<jbjnr> havefun. when are you back.
<jbjnr> and is aurianne away too
<simbergm> I mean everyone is on holiday tomorrow, it's the national day
<jbjnr> oh yes.
<simbergm> I'm staying around, just going to try not to work tomorrow, back on friday again
<simbergm> same with auriane I assume
<jbjnr> ok
<simbergm> fyi auriane = rori here on irc
<simbergm> good holidays?
heller has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
heller has joined #ste||ar
<jbjnr> simbergm: yes. we were in Ireland for two weeks. Rained almost every day and max temp was about 18deg, so it was not a heatwave!
<simbergm> I take that as a good thing?
<jbjnr> The rest of the family were glad about it, but I prefer the warm weather.
<Yorlik> I am running into a compile issue with the latest "stable":
<Yorlik> include\hpx\lcos\packaged_action.hpp(135): fatal error C1001: An internal error has occurred in the compiler.
<Yorlik> (compiler file 'msc1.cpp', line 1519)
<Yorlik> To work around this problem, try simplifying or changing the program near the locations listed above.
<Yorlik> Anyone encountered this too? Is it a known issue?
<Yorlik> I didn't have this on Linux
<Yorlik> Actually stable itself compiled
<zao> ICEs are mysterious beasts, no rhyme or reason.
<Yorlik> Its the user of HPX triggering this issue
<Yorlik> Triggered by this: hpx::cout << "Producer has returned ..." << hpx::endl;
<zao> A compiler should never crash, regardless of whether the code is legal or not. If it's supposedly legal code, you have more traction in filing bugs with the vendor, but a repro case that is "get HPX and Boost" is never going to get touched.
<jbjnr> just use std::cout
<zao> Cool people use creduce, but it's quite a bother on Windows I reckon.
<Yorlik> I'm not cool - just trying to get stuff done ;)
eschnett has quit [Quit: eschnett]
hkaiser has joined #ste||ar
diehlpk has joined #ste||ar
<Yorlik> Any idea how I could debug and fix this? https://gist.github.com/McKillroy/4ed6f9c56155282ec0bc46f4d2c7525c
<Yorlik> I'm hittinmg quite a wall here.
eschnett has joined #ste||ar
<hkaiser> Yorlik: use a different compiler to see the problem, msvc often crumbles on obscure syntax or other problems
<Yorlik> It seems the same problem with packaged_action.hpp is comng up also in another file
<Yorlik> I have no issues on Linux with Clang
<hkaiser> ok
<Yorlik> Its in the git you have access to if you feel like looking into it
<Yorlik> The HPX enabled version of the disruptor app
<Yorlik> The non HPX version and its tests work
diehlpk has quit [Ping timeout: 272 seconds]
<Yorlik> Another line of code in another unrelated sourcefile triggers it too:
<Yorlik> auto rFut = hpx::async<agns::game::lua::solstate::init_action>( mSolId );
<hkaiser> Yorlik: report it to Stephan (STL), he's very aggressive internally about ICE's
<hkaiser> I have never seen this particular ICE
<Yorlik> Which Stephan?
<K-ballo> I hear the latest update ICEs a lot, even randomly
<Yorlik> Argh
<Yorlik> Clang not working, cl ICEing - great
<Yorlik> My windows build is fudged now
<hkaiser> Yorlik: give him a preprocessed file, so he can reproduce the issue
<Yorlik> Arg - now I have to learn that as well .. lol
<hkaiser> use /P on the compilers command line, it will create a file *.i which is what you need
<Yorlik> Allright - thanks !
hkaiser has quit [Quit: bye]
<zao> Yorlik: To elaborate on what creduce does, it's a tool that take a preprocessed source file and runs it through test script to see if it's exhibiting a problem or is irrelevant. It then simplifies the source as much as it can via libclang and horrible Perl use, to "reduce" the test case to the bare minimal needed to exhibit the problem.
<zao> Small reduced repro cases are way easier than 500 megabytes of preprocessed C++ to troubleshoot :)
<Yorlik> IC - sound slick :)
<zao> But if you can't reduce it, a preprocessed source file with the flags used are better than nothing.
<zao> I did it a fair bit when we used to provide the Pathscale compilers.
<zao> That product is thankfully quite stale now, we only do GCC and Intel unless someone explicitly requires some other compiler.
<zao> Used to build _all_ software with the least horrible of GCC, Intel, Pathscale, Portland.
hkaiser has joined #ste||ar
aserio has joined #ste||ar
diehlpk has joined #ste||ar
<Yorlik> I kinda don't really know what to do now. Seems I have to find a stable commit cl can digest
<hkaiser> or find a cl that is usable ;-)
<hkaiser> simplify the code, that might help as well
<hkaiser> try to reduce your code to a minimal failing case
diehlpk has quit [Ping timeout: 245 seconds]
daissgr has quit [Read error: Connection reset by peer]
aserio1 has joined #ste||ar
aserio has quit [Ping timeout: 264 seconds]
aserio1 is now known as aserio
aserio has quit [Read error: Connection reset by peer]
aserio has joined #ste||ar
aserio has quit [Ping timeout: 244 seconds]
aserio has joined #ste||ar
<rori> /quit ye!
rori has quit [Quit: bye!]
diehlpk has joined #ste||ar
diehlpk has quit [Remote host closed the connection]
diehlpk has joined #ste||ar
<diehlpk> aserio, hkaiser Do you want to do more changes for the BOF or can I upload the first draft?
<aserio> diehlpk: Please upload the first draft
<diehlpk> aserio, I need a pic from you
<aserio> I will send one after the meeting I am in
diehlpk has quit [Ping timeout: 245 seconds]
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar
<nikunj97> hkaiser: yt?
<hkaiser> here
diehlpk has joined #ste||ar
<nikunj97> I was thinking if we really need anything other than 16 core results
<nikunj97> We might want to show scaling of overheads
<diehlpk> hkaiser, aserio Should I add all attendes as session leaders or only aserio and me?
<nikunj97> but other than that, I don't see reasons for including them
<hkaiser> nod
<hkaiser> we should show how overheads depend on the number of cores used
<nikunj97> nod
<nikunj97> I'll keep the error rate vs execution time only for 16 cores then
<nikunj97> that way, it will be easier for me to show the bevior
<hkaiser> ok
<nikunj97> and also write things up clearly
<nikunj97> great! I'll go ahead with this then. Also, I got an email from Kevin
<nikunj97> he's approved it
<nikunj97> now NERSC has some work to do, then I should have access to cori
<hkaiser> nikunj97: perfect!
<diehlpk> hkaiser, yet?
<hkaiser> here
<diehlpk> Should I add all people from the previous panel as session leaders to the BOF?
<diehlpk> Or just Adrian and me?
<diehlpk> I am nit sure what they mean by session leader
<hkaiser> diehlpk: session leader would be you, you may add the others as planned speakers/contributors
<diehlpk> Ok, I already did it
<hkaiser> ok
<diehlpk> I have me and Adrian as session leaders
<hkaiser> ok, fine
<diehlpk> Adrian is my backup
<hkaiser> he is everybodies backup ;-)
<diehlpk> Yeah, ok I submitted the first draft
<hkaiser> thanks!
<diehlpk> We still could make changes until midnight, so please let mek now if you like to change things
<hkaiser> ok, I'll have a look
diehlpk has quit [Ping timeout: 268 seconds]
eschnett has quit [Quit: eschnett]
aserio has quit [Quit: aserio]
eschnett has joined #ste||ar
Yorlik has quit [Read error: Connection reset by peer]
quaz0r has quit [Quit: reboot]