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/
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
<github> [hpx] hkaiser opened pull request #2849: Unify serialization of non-default-constructable types (master...fixing_2846) https://git.io/v5erw
vamatya has quit [Ping timeout: 255 seconds]
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/v5eoM
<github> hpx/master d4ec908 Hartmut Kaiser: Fixing tests
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/v5eKM
<github> hpx/master e157e1c Hartmut Kaiser: Disable HDF5 examples as cmake v3.9 broke HDF5 detection
eschnett has joined #ste||ar
<zao> CMake breaking things? Amazing.
hkaiser has quit [Quit: bye]
mcopik has quit [Ping timeout: 240 seconds]
K-ballo has quit [Quit: K-ballo]
vamatya has joined #ste||ar
vamatya has quit [Ping timeout: 248 seconds]
jaafar_ has quit [Ping timeout: 276 seconds]
bikineev has joined #ste||ar
patg has quit [Quit: This computer has gone to sleep]
bikineev has quit [Remote host closed the connection]
pree has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
mcopik has joined #ste||ar
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
hkaiser has joined #ste||ar
bikineev has quit [Read error: Connection reset by peer]
bikineev has joined #ste||ar
<pree> Hi all. what minimum c++ version should hpx support ? Because I'm using some c++17 things like std::conjuction's etc. what if the user has compiler which supports upto c++11
<pree> the build gets failed ?/
<hkaiser> pree: in unguarded code, please stick to c++11
<pree> Okay. Thanks
<pree> unguared code ??
<pree> hkaiser ^^
<hkaiser> if you want to use features of c++14 or 17 place those code snippets inside #ifdef/#endif
<hkaiser> best is to stick with c++11, though
<pree> Okay I will try to stick with c++11.
<pree> and some ifdef's if want to
<hkaiser> pree: using #ifdef means you need two alternatives of the code, c++11 and a newer version - I think you shouldn't even think of this at this point
<pree> okay
<hkaiser> pree: I doubt you have a use case where you can't get away with c++11
<pree> nothing special just for some code readabilty though
<hkaiser> pree: having two versions does not increase readability, usually
<pree> Okay I will stick to c++11,
<pree> thanks
<pree> I thought it will give readability without alternatives, now I'm okay
K-ballo has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree has quit [Ping timeout: 246 seconds]
Rodario has joined #ste||ar
<Rodario> Hi, smone here i could annoy with a question regarding the build process of hpx relly fast?
pree has joined #ste||ar
<zao> I'm going to conjure up breakfast, but you can always try :)
<hkaiser> zao: breakfast?
<jbjnr> 14:39 here. start the BBQ soon!
<hkaiser> Rodario: ust ask, somebody will answer
<zao> hkaiser: Sat up a bit too long last night to try to tease out my CPU bugs with HPX :)
<zao> Didn't learn much, except that if you build it under the Windows Subsystem for Linux, compiler processes apparently do not want to touch swap.
<hkaiser> zao: almost like in the US, something wrong - the russians!
<hkaiser> zao: for you - something wrong - HPX!
<hkaiser> ;)
<Rodario> allrighty, currently my build dies with error msg referring to undefined refrences. I realized then that while using cmake i get a msg saying: "failed to create symbolic link <root_path>/hpx_main/hpx/build/share/hpx': No such file or directory"
<zao> It's a good load on a system :)
<hkaiser> Rodario: that one is benign, you can safely ignore it
<hkaiser> note to self - find a way to avoid this warning
<Rodario> huh ... bummer. I hoped this was the problem crashing the build :D
<zao> Do you have some background on what kind of environment you're building in, and what are the missing symbols and for what binary?
<Rodario> the first one is: numa_domains.cpp:(.text+0x9c): undefined reference to `hpx::get_runtime()
<github> [hpx] hkaiser pushed 2 new commits to master: https://git.io/v5vvk
<github> hpx/master f9ec761 Denis Blank: Remove constexpr from a void function
<github> hpx/master 6609147 Hartmut Kaiser: Merge pull request #2840 from Naios/constexpr_void...
<Rodario> I put the informative part of the error on pastebin: https://pastebin.com/PeF9Kq6G
<zao> What distro/compiler version is this, and is it HPX master or a released version?
<zao> (I guess the HPX version is answered by the soname)
<Rodario> I am using gcc 7.1 and fedora 26 with the newest hpx master. When using cmake i define HPX_WITH_CXX14=ON.
<hkaiser> Rodario: I have never seen this problem before
<hkaiser> Rodario: is that a fresh build ?
<hkaiser> i.e. in an empty directory?
<hkaiser> no other HPX versions installed?
<Rodario> yes
<Rodario> and yes
<Rodario> or does fedora come with hpx?
<hkaiser> Rodario: please give us the full logs, then - cmake, make
<hkaiser> Rodario: don't think so
<hkaiser> preferrably with make VERBOSE=1
<Rodario> me neither :D full logs mean my full error messages or does hpx write logs after crashing the build?
<hkaiser> Rodario: full build logs including your command lines
<hkaiser> whatever cmake and make print while building
<Rodario> ok, in pastebin again?
<hkaiser> all of it
<hkaiser> yes, please
<Rodario> on it
bikineev has joined #ste||ar
<zao> I haven't built on F26 release yet, just the beta.
<Rodario> this is the cmake output: https://pastebin.com/LePCstBz
<Rodario> im currently building
bikineev has quit [Remote host closed the connection]
<Rodario> the error occurs after: Linking CXX shared library ../lib/libhpx.so
<Rodario> the verbose output of make -j3: https://pastebin.com/9uYXmVr8
<hkaiser> rodario can you give me the output of cmake as well, please?
<hkaiser> including command line, pls
<hkaiser> also, you did not give me the full build logs, nor the errors generated
<zao> (did you know that machines don't boot without RAM?)
<Rodario> command line for cmake would be: cmake -DHPX_WITH_CXX14=ON VERBOSE=1 ..
<Rodario> or do I miss something here
<hkaiser> VERBOSE=1 is needed for make only, the HPX_WITH_CXX14=On goes onto the cmake command line
<hkaiser> zao: they tend to squirm without RAM, yah
<Rodario> allright, i posted the verbose output, the error msg. that comes after it is bigger than my terminal can scroll. How to post all of that? I am rather novice concerning terminal usage ...
vamatya has joined #ste||ar
<zao> Typically, you can glue on 2>&1
<zao> Typically, you can glue on 2>&1 | tee -a some-log-file
<zao> (silly enter key)
<zao> Merges stderr into stdout, and pipes it into tee, which shows the output to screen and also appends it into the file mentioned.
<zao> You can then drag the file into a gist, or otherwise share it.
<Rodario> Cool stuff. Now to get through those 16 764 lines of error...
<zao> hkaiser: ^
vamatya has quit [Ping timeout: 276 seconds]
<Rodario> looks like gist does not want to read this many lines either. it stops after 1883
<zao> There's a link up top to "view the full file"
<Rodario> ah
<zao> I don't have any immediate ideas. Rather odd.
<Rodario> did i miss to post anything
<Rodario> ?
<zao> Should be everything, I believe.
<zao> I'
<zao> I'm setting up a F26 VM to see if I can tease this out, no idea what hkaiser is up to :)
<hkaiser> Rodario: sorry - need to leave now - will try to look later
<K-ballo> shows only linking stage
<zao> If possible, a log from a clean build might be helpful.
<K-ballo> I'm mostly curious on the flags used to compile the sources, just one invocation would be enough
<github> [hpx] hkaiser force-pushed fixing_2847 from d801db8 to 3778f07: https://git.io/v5ect
<github> hpx/fixing_2847 3778f07 Hartmut Kaiser: Enable returning future<R> from actions where R is not default-constructible...
<zao> F26 is on CMake 3.9.0, btw.
<Rodario> which is a problem?
<K-ballo> looks sane
<zao> Nah, just an observation.
<zao> [ 12%] Linking CXX shared library ../lib/libhpx.so
<zao> [ 35%] Built target hpx
<zao> libhpx seems to have built on my box.
<K-ballo> Rodario: could you find out which flags you are using?
<Rodario> flags for cmake?
<Rodario> like: -DHPX_WITH_CXX14=ON
<Rodario> ?
<zao> For a /usr/bin/c++ invocation.
<zao> Like the one I linked above, with make VERBOSE=1
<K-ballo> you may need to force a rebuild, you may pass -B to make
<zao> or delete some object file, like src/CMakeFiles/hpx.dir/util/batch_environment.cpp.o
<K-ballo> much easier to do that, yes
<Rodario> an then use make with VERBOSE=1 again?
<zao> Right.
<Rodario> make VERBOSE=1 -j3
<Rodario> but wasnt this the way I created the last gist?
<K-ballo> for the last gist when make was run all translation units had already been compiled, so it only shows the attempt to link all those object files into a library
<K-ballo> deleting one of those object files will result in make running the corresponding compilation step for that translation unit, and revealing the flags used for compilation
<Rodario> but this does not have the output compiler flags I see at zaos post
<K-ballo> you did not remove batch_environment.cpp.o as zao suggested
<zao> err, bah
<zao> Sliced it out.
<zao> compile + link
<Rodario> oh, wow
<zao> I see that you've got your Boost in /usr/local. Did you build that yourself?
<Rodario> yes, I did.
<Rodario> but cmake detected it
<Rodario> did I build the library wrong?
<zao> I'm leaning toward that either the system headers are found for Boost, if present, or your Boost may be built wrong.
<zao> Not sure of the required flags, but I build mine 'cxxflags=-std=c++14' when building HPX in C++14 mode.
<Rodario> ooooh
<zao> Boost 1.61 is a bit old too, so not sure if it's going to work overly well with GCC 7
<zao> Let's see if I can break mine :)
<Rodario> that would suggest i remove boost from my mashine and build it with c++14 flag?
pree has quit [Ping timeout: 246 seconds]
<zao> stock 1.61.0 build linked in fine too, so not that either.
<zao> Which commit of HPX are you on? `git describe` for me is 0.9.11-rc1-4123-g6609147484
<Rodario> 0.9.11-rc1-4123-g6609147484
<Rodario> jup
<Rodario> I also build f26 from scratch just a week ago
<Rodario> do you also get the error about the symbolic link that could not be set?
<Rodario> during cmake
<zao> Yeah, that's universal.
<zao> One of those cosmetic things we haven't gotten around nuking yet.
<Rodario> ok
<zao> I'm quite at a lack of ideas now... have you done anything weird with this installation at all? Is it up to date?
<zao> Also, "build f26 from scratch", was that an actual build of the OS, or just an installation?
<K-ballo> we all get the symbolic link error.. where does that come from?
<Rodario> uhm, i installed it
<Rodario> not build, thats right
<Rodario> I think I will remove all components and install a new, will report back ;)
<Rodario> thanks for the great help tho
<K-ballo> please do report back
<K-ballo> now I'm curious
pree has joined #ste||ar
parsa has joined #ste||ar
<github> [hpx] hkaiser force-pushed fixing_2847 from 3778f07 to e4c1c0a: https://git.io/v5ect
<github> hpx/fixing_2847 e4c1c0a Hartmut Kaiser: Enable returning future<R> from actions where R is not default-constructible...
bikineev has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
<7JTAB2CB3> [hpx] hkaiser created fixing_2848 (+1 new commit): https://git.io/v5vOs
<7JTAB2CB3> hpx/fixing_2848 18cec56 Hartmut Kaiser: Allowing for non-default-constructible component types...
parsa has joined #ste||ar
diehlpk has joined #ste||ar
diehlpk has quit [Ping timeout: 248 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
patg has joined #ste||ar
patg is now known as Guest96345
<github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://git.io/v5vc7
<github> hpx/gh-pages 79c3d51 StellarBot: Updating docs
jaafar_ has joined #ste||ar
pree has quit [Ping timeout: 246 seconds]
pree has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
diehlpk has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree has quit [Ping timeout: 246 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
bikineev has joined #ste||ar
pree has joined #ste||ar
parsa has joined #ste||ar
diehlpk has quit [Remote host closed the connection]
pree has quit [Quit: AaBbCc]
Rodario has quit [Quit: Leaving.]
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/v5vzX
<github> hpx/master 0a98734 Hartmut Kaiser: Fixing more tests
<K-ballo> what happened to the huge variant of messages that was part of the addressing server (or related)?
<github> [hpx] K-ballo force-pushed pack-short-circuit from 37dfaa3 to fde9b6b: https://git.io/v762b
<github> hpx/pack-short-circuit fde9b6b Agustin K-ballo Berge: Short-circuit all_of/any_of/none_of instantiations
<hkaiser> K-ballo: not sure what you're referring to
<K-ballo> hkaiser: a long time ago there were a `request` and a `response` boost::variants of tuples or something like that
<hkaiser> K-ballo: ahh yes, those are gone
<hkaiser> K-ballo: I think boost::variant is only used for some of the spirit parsers nowadays
<K-ballo> yes, some policy mapping or something
<hkaiser> just attributes for Qi alternatives
<hkaiser> iirc
<github> [hpx] hkaiser pushed 1 new commit to fixing_2847: https://git.io/v5vVi
<github> hpx/fixing_2847 7d28e21 Hartmut Kaiser: Making future serialization symmetric even for non-default-constructible value types