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/
<zao> K-ballo: top said clang build used /usr/bin/ld so it must just be a nicer .o going in
<zao> hkaiser: soumds nice :)
vamatya has quit [Ping timeout: 268 seconds]
EverYoung has quit [Ping timeout: 246 seconds]
http_GK1wmSU has joined #ste||ar
http_GK1wmSU has left #ste||ar [#ste||ar]
hkaiser has quit [Quit: bye]
parsa has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
pagrubel has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
vamatya has joined #ste||ar
vamatya has quit [Ping timeout: 276 seconds]
david_pfander has joined #ste||ar
ajaivgeorge has quit [Ping timeout: 255 seconds]
ajaivgeorge has joined #ste||ar
ajaivgeorge has quit [Ping timeout: 246 seconds]
david_pfander has quit [Ping timeout: 276 seconds]
david_pfander has joined #ste||ar
ajaivgeorge has joined #ste||ar
bikineev has joined #ste||ar
<github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://git.io/v7g2c
<github> hpx/gh-pages 29600eb StellarBot: Updating docs
bikineev has quit [Remote host closed the connection]
david_pfander has quit [Quit: david_pfander]
david_pfander has joined #ste||ar
pagrubel has quit [Remote host closed the connection]
pagrubel has joined #ste||ar
mcopik has joined #ste||ar
david_pfander1 has joined #ste||ar
david_pfander has quit [Ping timeout: 260 seconds]
david_pfander1 is now known as david_pfander
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
<github> [hpx] hkaiser force-pushed fixing_2796 from 71bc36d to 798b7e7: https://git.io/v7zUF
<github> hpx/fixing_2796 798b7e7 Hartmut Kaiser: Making sure future::wait_for et.al. work properly for action results...
ajaivgeorge has quit [Ping timeout: 240 seconds]
eschnett has quit [Quit: eschnett]
hkaiser has quit [Read error: Connection reset by peer]
hkaiser has joined #ste||ar
pree_ has joined #ste||ar
pree_ has quit [Read error: Connection reset by peer]
<jbjnr> hkaiser: you broke all my stuff!
pree_ has joined #ste||ar
pree_ has quit [Read error: Connection reset by peer]
eschnett has joined #ste||ar
pree_ has joined #ste||ar
parsa[w] has joined #ste||ar
<zao> Woo-hoo!
<jbjnr> zao: has found a portal to another dimension
<K-ballo> not Xen I hope
<zao> This dimension sucks, there's code like `float** s=new(float*[q->L+2]);` and `inline float log2(float x) {return (x<=0? (float)(-100000):1.442695041*log(x));}` in it.
pree_ has quit [Read error: Connection reset by peer]
<hkaiser> jbjnr: it gets worse, more incomming ;)
<hkaiser> I'll tell you how to fix things
<jbjnr> Oh no. I'm in conflict hell already. I've been working on the thread pools too.
pagrubel has quit [Ping timeout: 246 seconds]
<hkaiser> uhh
<hkaiser> not good
<jbjnr> Do what you need to do, then let me know so I can work on my stuff
<hkaiser> well, you resolve your conflicts and push, I'll resolve mine afterwards
<hkaiser> jbjnr: I'm sorry
<jbjnr> I just removed your thread_pool changes from my branch. wasted too long on getting things running that I decided to just get on with my tests etc
<hkaiser> trying to clean up things...
<jbjnr> I am happy you are cleaning up.
<hkaiser> jbjnr: if you gave me access to your code I could fix things for you
<hkaiser> jbjnr: ok, then push your changes at least
<hkaiser> I can resolve conflicts and give it back to you
<jbjnr> my changes to simple resource manager are not stable. I am testing new featers in schedulers
<jbjnr> ^features
<jbjnr> but I need the schedulers to link. so removing the template stuff from the header broke it
<hkaiser> jbjnr: I know, that's what I'm fixing right now
<jbjnr> ok
<hkaiser> but I also renamed things, so this will cause problems
<jbjnr> thanks
pree_ has joined #ste||ar
<hkaiser> jbjnr: so things link again for me, should I push or should I wait for you?
pree_ has quit [Read error: Connection reset by peer]
<jbjnr> please push. I will pick upi your stuff later. I am going through thread numbering issues
<hkaiser> jbjnr: give me the example exposing the thread numbering problems, I can fix those quickly
<github> [hpx] hkaiser pushed 1 new commit to resource_partitioner: https://git.io/v7gxI
<github> hpx/resource_partitioner 406be02 Hartmut Kaiser: Renaming thread pool classes...
<hkaiser> jbjnr: good luck ^^
<jbjnr> I'm just using the simple resource_manager example with different command line options etc
<jbjnr> but the thread numbering is something that has got broeken and I wanted to go through it all carefully and check it
<hkaiser> jbjnr: it was broken before I changed things, though
<hkaiser> how do you know there are numbering problems?
<jbjnr> so it seems, but I am not sure how that happened and I don't trust anything now
<hkaiser> so how do you know there are numbering problems?
<jbjnr> segfaults etc
<hkaiser> what command line options cause this?
<hkaiser> the example runs for me :/
<github> [hpx] hkaiser force-pushed resource_partitioner from 406be02 to a481fa8: https://git.io/v7lfK
<github> hpx/resource_partitioner a481fa8 Hartmut Kaiser: Renaming thread pool classes...
hkaiser has quit [Quit: bye]
pree_ has joined #ste||ar
<heller> it would be so nice if the complete virtual address space would be automatically mapped to the NIC
<heller> without special registration
aserio has joined #ste||ar
hkaiser has joined #ste||ar
aserio has quit [Ping timeout: 246 seconds]
aserio has joined #ste||ar
<aserio> heller, jbjnr, wash[m]: Will you be joining us today?
<jbjnr> can do, not much to talk about though
<jbjnr> hkaiser: thanks, branch compiling again now
<jbjnr> hkaiser: please warn me if you are planning any more changes like that :)
<hkaiser> jbjnr: great
<hkaiser> jbjnr: will do
<hkaiser> sorry again
<heller> aserio: forgot
<heller> aserio: give me a secon
<jbjnr> I'm mostly working on the custom scheduler, but keep making tweaks to the thread pools and rp classes and have lots of small uncommitted bits'n'pieces that keep changing
<aserio> david_pfander: ^^
<hkaiser> jbjnr: just commit what you have, please
<jbjnr> no! it's unfinished work
<hkaiser> hmmm
<hkaiser> jbjnr: I need to implement the MPI/HPX bridge stuff soon, that's my driving use case
<jbjnr> has the meeting started?
<heller> looks so
<wash[m]> Yes
pree_ has quit [Read error: Connection reset by peer]
<aserio> wash[m]: how is the paper comming?
pree_ has joined #ste||ar
<wash[m]> Good. Should have another draft tmrw
<wash[m]> Err, today even
<wash[m]> aserio: wait what time was the meeting, did I just miss it? Just realized it wasnt on my calendar
bikineev has joined #ste||ar
<zao> I see that the lab is up and running again, yay.
pree__ has joined #ste||ar
pree_ has quit [Ping timeout: 260 seconds]
<aserio> wash[m]: lol, good to hear
bibek_desktop has quit [Quit: Leaving]
pree__ has quit [Ping timeout: 260 seconds]
aserio has quit [Ping timeout: 246 seconds]
bibek_desktop has joined #ste||ar
aserio has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree__ has joined #ste||ar
pree__ has quit [Ping timeout: 260 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
pree__ has joined #ste||ar
pree__ has quit [Read error: Connection reset by peer]
pree__ has joined #ste||ar
pree__ has quit [Ping timeout: 260 seconds]
Vir has quit [Ping timeout: 240 seconds]
aserio has quit [Ping timeout: 258 seconds]
pree__ has joined #ste||ar
pree__ has quit [Ping timeout: 260 seconds]
mcopik has quit [Ping timeout: 240 seconds]
pree_ has joined #ste||ar
pree_ has quit [Ping timeout: 240 seconds]
EverYoung has quit [Ping timeout: 246 seconds]
EverYoung has joined #ste||ar
aserio has joined #ste||ar
pree_ has joined #ste||ar
<github> [hpx] hkaiser created partitioned_vector (+1 new commit): https://git.io/v72uL
<github> hpx/partitioned_vector 307c280 Hartmut Kaiser: Speeding up compile/link time for applications using partitioned_vector...
<github> [hpx] hkaiser closed pull request #2797: Making sure future::wait_for et.al. work properly for action results (master...fixing_2796) https://git.io/v7u9A
<hkaiser> zao: that should lessen the pressure on the compiler and the linker^^, let's see
<zao> Ooh.
<zao> I'm conveniently on my way home soon.
pree_ has quit [Read error: Connection reset by peer]
<zao> debug build of partitioned_vector_transform_scan.cpp with gcc - 5990MB compile.
<zao> 1GB link.
<zao> (VIRT)
<zao> On that `partitioned_vector` branch.
pagrubel has joined #ste||ar
<zao> So almost a gigabyte less memory to compile, dito to link.
<zao> Measurements may be a bit off toward the end, as I watched usage in top.
<zao> Now going home :)
<hkaiser> so it helped
<zao> Oh wait, it failed to link...
aserio has quit [Quit: aserio]
<hkaiser> uhh
<hkaiser> links for me - but I'll see what CI does
<zao> Just ran a build in an existing tree, so I may need to rebuild stuff more?
<zao> It built something like 182 targets in total.
<hkaiser> strange, will look
pree_ has joined #ste||ar
<zao> Only odd settings I have is -DHPX_WITH_CXX14, on GCC 6.3.0.
<zao> Spun up a full build in a distinct tree, anyway.
<hkaiser> thanks
<zao> Bought a bigger SSD yesterday, so now I can fit more than one HPX build on disk :P
<hkaiser> lol
<hkaiser> we'll fill THAT up as well ;)
zbyerly_ has joined #ste||ar
<zao> Fresh build tree.
<hkaiser> zao: just this failue or do all PV tests fail to link?
<zao> Don't know, build aborts at that poin.
<hkaiser> ok, so this is justthe first failure
<zao> partitioned_vector_2201_test passes f.ex.
<zao> *poof*
<hkaiser> interesting - will investigate later tonight
pree_ has quit [Ping timeout: 246 seconds]
mcopik has joined #ste||ar
<heller> oh gosh will I cry if both github and my computer looses my thesis
pree_ has joined #ste||ar
zbyerly__ has joined #ste||ar
zbyerly_ has quit [Ping timeout: 260 seconds]
<pree_> parsa[w] : yt ?
pree_ has quit [Ping timeout: 258 seconds]
<jbjnr> heller: loose is what your belt is when your pants fall down. lose is what github will do to your thesis if you can't has english properly
pree_ has joined #ste||ar
<zao> heller: Om nom nom.
<zao> Fun fact, I encrypted the disk image for my thesis project code and assets, due to NDA with the company I worked with.
<zao> I lost the password <_<
<heller> lol
<heller> jbjnr: I will loose my patience soon
pree_ has quit [Ping timeout: 260 seconds]
zbyerly__ has quit [Ping timeout: 255 seconds]
<heller> especially with the clusterfuck they call latex image placement algorithm
<zao> [!htbffsaaargh]
<zao> \flushallfiguresgoshdarnit
pree_ has joined #ste||ar
<zao> partitioned_vector_iter_test also fails to link, thus far.
<zao> (running ninja -k0)
<zao> That has partitioned_vector<int> and partitioned_vector<double> missing.
Vir has joined #ste||ar
pree_ has quit [Ping timeout: 260 seconds]
<zao> And more. Yay.
pree_ has joined #ste||ar
pree_ has quit [Ping timeout: 260 seconds]
pree_ has joined #ste||ar
zbyerly_ has joined #ste||ar
aserio has joined #ste||ar
pagrubel is now known as patg[[w]]
jfbastien has joined #ste||ar
hkaiser has quit [Quit: bye]
pree_ has quit [Ping timeout: 276 seconds]
pree_ has joined #ste||ar
mars0000 has joined #ste||ar
hkaiser has joined #ste||ar
pree_ has quit [Ping timeout: 260 seconds]
pree_ has joined #ste||ar
aserio has quit [Quit: aserio]
pree_ has quit [Ping timeout: 240 seconds]
<jbjnr> zao: is that true that you lost the password? to your thesis work?
bikineev has joined #ste||ar
<zao> Sure is!
<zao> Redid everything from scratch with some older assets I could get from our defunct FTP.
<jbjnr> that's really shocking. in several ways
<zao> jbjnr: Company I did thesis for bankrupt and I took a break from studies, doing consulting in webdev.
<zao> In the time after that I didn't care overly much about finishing it as I had other more urgent things like keeping a roof and food :)
<zao> So when I got back to it, I found a disk image and a memory that I had a very good passphrase indeed :)
<zao> At that point, I didn't have any implementation at all, and very little notes on literature and whatnot.
<zao> So I fretted about it and did it from scratch over the next five years or so :D
mars0000 has quit [Quit: mars0000]
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 240 seconds]
EverYoun_ has quit [Ping timeout: 246 seconds]
EverYoung has joined #ste||ar
<zao> jbjnr: It'd be fun if I ever applied for a PhD programme... "so it says here that you did began your master thesis in 2008 and completed it in 2015?"
bikineev has quit [Remote host closed the connection]
eschnett has quit [Quit: eschnett]
mcopik has quit [Ping timeout: 258 seconds]
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
<zao> /usr/home/zao/stellar/hpx/hpx/runtime/threads/coroutines/detail/posix_utility.hpp:91:43: error: use of undeclared identifier 'MAP_NORESERVE'
<zao> MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE,
<zao> Is that something we've sucked in from Boost? Breaks out of the box on FreeBSD :)
<zao> Guess there's some knob somewhere I keep forgettin.
<zao> Ah, MAP_NORESERVE is removed in FreeBSD 11.x, and never did anything in previous versions :D
<zao> /usr/home/zao/stellar/hpx/hpx/runtime/threads/coroutines/detail/context_impl.hpp:143:2: error: No default_context_impl available for this system
<zao> Bleh.
<zao> Going for GENERIC, see how it fares.
<zao> (why do I keep insisting on this OS?)
<zao> src/CMakeFiles/hpx.dir/exception.cpp.o: In function `hpx::detail::get_execution_environment()':
<zao> /usr/home/zao/stellar/hpx/src/exception.cpp:(.text+0x12bc): undefined reference to `environ'
<zao> Thought we'd solved that one before... guess it snuck back in.
<zao> While linking libhpxd.so
<hkaiser> zao is muttering to himself again ...
<hkaiser> anything I can do to help?
<zao> Not sure yet, just seeing how far I get on FreeBSD on the other spare machine.
<zao> Thus far I've gotten around things, but this I'm not quite sure.
<zao> exception.cpp seems to want to access the environment, but the symbol 'environ' is not defined in a shared library, and the link bails out.
<hkaiser> give me a sec
<zao> -z,defs mutters something about not allowing undefined symbols in object files, but allowing it in shared libraries.
<zao> I assume that means inputs to the link process and not the output.
<zao> Apple platforms has some function to acquire the environ pointer, judging by our sources.
<hkaiser> nod
<hkaiser> how is this done on freebsd?
<hkaiser> how can I access the global environment?
<hkaiser> all of it, that is
<zao> environ is the way, but it's only present in executables.
<hkaiser> zao: it seemed to have worked before on freebsd, at least it's mentioned in the pp constant there
<zao> Which because we disallow undefined symbols, breaks the build of the HPX shared library.
<zao> Yeah, saw that further down.
<zao> Not sure when we started with -z,defs
<zao> Or they've removed it in later libc.so versions on FBSD.
<zao> WebKit just gave up and removed their use completely.
<hkaiser> zao: https://www.freebsd.org/cgi/man.cgi?environ(7) does not say anything about only available in executables
<zao> There's nothing intrinsically _wrong_ about using it in a shared objcet.
<zao> It's because we prevent undefined symbols it goes bad.
<zao> It'll (assumedly) get resolved at dynamic link time, haven't tested.
<hkaiser> k
<hkaiser> we could simply define the variable ourselves for freebsd
<hkaiser> and initialize it from within hpx::init or so
<zao> My shared library (if allowed to have undefined externals) loads and runs properly in a test binary, so the linker resolves it eventually.
<zao> hkaiser: Hmm, interesting.
<hkaiser> zao: try adding 'HPX_EXPORT char** freebsd_environ = nullptr;' and assign environ to it from here: https://github.com/STEllAR-GROUP/hpx/blob/master/src/util/find_prefix.cpp#L45
<zao> Regarding my linker errors on the partitioned_vector branch, here's my uncurated build log - https://gist.github.com/zao/8e7192d0d50b42abb5093425bec72e36
<zao> Assuming that Gists handle a megabyte of output :)
<hkaiser> zao: yah something goes horribly wrong for you with the PV
<hkaiser> no idea yet what actually
<hkaiser> all of the missing symbols should be exported from the partitioned_vector.so