hkaiser changed the topic of #ste||ar to: The topic is '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/
K-ballo has quit [Ping timeout: 268 seconds]
K-ballo has joined #ste||ar
primef has quit [Ping timeout: 260 seconds]
hkaiser has quit [Ping timeout: 260 seconds]
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
primef has joined #ste||ar
hkaiser has joined #ste||ar
adi_ahj has joined #ste||ar
primef has quit [Ping timeout: 268 seconds]
nikunj has joined #ste||ar
<Yorlik> Is there a nown problem with performance counters in 1.4.0 7 windows? My build broke switching counters on.
<Yorlik> /s7///g
<Yorlik> crap - seems evem my regex is borked today - :)
adi_ahj has left #ste||ar [#ste||ar]
hkaiser has quit [Ping timeout: 260 seconds]
hkaiser has joined #ste||ar
<hkaiser> Yorlik: what broke?
<hkaiser> Yorlik: I have the pdb issue reproduced, btw
<hkaiser> need to find a fix now
<Yorlik> I'm double checking right now - I'll come up with more details soonish. I was some unknown symbol
nikunj has quit [Quit: Leaving]
primef has joined #ste||ar
primef has quit [Ping timeout: 268 seconds]
primef has joined #ste||ar
<Yorlik> hkaiser: -DHPX_WITH_IO_COUNTERS=ON busts the build (windows, debug tested):
<Yorlik> C:\__A\AC_DOWNLOAD\hpx\1.4.0\components\performance_counters\io\src\io_counters.cpp(82): error C2065: 'pid_t': undeclared identifier
<Yorlik> C:\__A\AC_DOWNLOAD\hpx\1.4.0\components\performance_counters\io\src\io_counters.cpp(82): error C2146: syntax error: missing ';' before identifier 'pid'
<hkaiser> Yorlik: that's anew one
<hkaiser> shouldn
<hkaiser> t be too hard to fix, however - looks like a missing #include
<Yorlik> All the other counters work
<Yorlik> The error message is pretty clear I think.
<Yorlik> I just don't understand why this passes on a major release
<Yorlik> Honestly - major releases (full minor version) imo should be fully tested on windows too.
<Yorlik> We are in the end goinf to use Lunux, but for development using windows is just easier.
<hkaiser> Yorlik: nobody builds with those counters enabled, apparently
<Yorlik> Your CD should do it
<hkaiser> indeed
<hkaiser> I wasn't even aware that this option is disabled by default :/
<Yorlik> Its a systematic CD/CI error. Should be easy to fix.
<hkaiser> Yorlik: 1.4.0 was a very large release, so please bear with us
<Yorlik> I see.
<Yorlik> Do you have some sort of release policy written somewhere? I am sticking with minor full numbers like 1.4.0 in the naive hope they are like really stable
<Yorlik> Also a separate bugfix branch for major releases might make sense.
<hkaiser> yes we're discussing a V1.4.1
<Yorlik> Not sure whats the best way to ensure stability while developing the cutting edge
adi_ahj has joined #ste||ar
<Yorlik> You could add my last two issues to the list.
<Yorlik> Both are on 1.4.0
<hkaiser> once we fix it we can discuss adding it, yes
<Yorlik> #4334 #4336 and the broken examples
<hkaiser> Yorlik: I think the io-counters never worked on windows - those are linux specific
<Yorlik> IC. Maybe we need a sort of extended build doc.
<Yorlik> I understand that would be quite a task - but the build options are a beast already.
<hkaiser> Yorlik: it should be possible to implement equivalent counters on windows, however - if you needed them
adi_ahj has quit [Ping timeout: 260 seconds]
<Yorlik> Not yet. I was just enabling all of them, because I started using counters
<Yorlik> But not really the IO ones
adi_ahj has joined #ste||ar
adi_ahj has quit [Client Quit]
<Yorlik> I am getting the following runtime error:
<Yorlik> {what}: unknown counter type /threads/time/average: HPX(bad_parameter)
<Yorlik> In C++ I declare it like this:
<Yorlik> hpx::performance_counters::performance_counter average_thread_counter ("/threads{locality#0/total}/time/average");
<Yorlik> I think it used to work if I'm not mistaken Did something change here or what would be the mistake?
<Yorlik> NVM - seems i missed an HPX build setting
<primef> hkaiser: I was wondering, if there is some (for now easy) issue I can work on in my spare time. As I would like to contribute to the project if time permits.
<hkaiser> primef: looks through the tickets
<primef> Alright, I will. Thanks!
<hkaiser> primef: what would be interesting to you?
<primef> hkaiser: Actually, I don't have any fixed interest. I could for example look into more build related topics or more source itself. Bare in mind, I'm quite new to HPX.
<hkaiser> sure, well anything goes
<hkaiser> we try to document open issues in the tickets, but sometimes there is more
<hkaiser> you might want to bring it up during the week here
<hkaiser> others might be responding
<primef> Alright, then I'll do that during the week. In the meantime I will give a look at the open tickets. Thanks
primef has quit [Ping timeout: 268 seconds]