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/
RostamLog has joined #ste||ar
<parsa[w]> hkaiser: i think i messed up your changes, don't pull/push for a few minutes
<parsa[w]> :d/i think/
eschnett has joined #ste||ar
<hkaiser> k
<parsa[w]> hkaiser: fixed. i lost some changes when i did resets but cross product works
<hkaiser> thanks
<hkaiser> aserio: brjsp got his first PR merged
zbyerly_ has quit [Ping timeout: 252 seconds]
diehlpk_work has quit [Quit: Leaving]
<hkaiser> parsa[w]: thanks, looks good
<jbjnr> how long before we can use if constexpr in our hpx code?
<jbjnr> do we know brjsp? seems like a very useful addition to the team if he makes more contributions!
<hkaiser> jbjnr: you can use HPX_CONSTEXPR, we have not made a decision on this yet
<hkaiser> there are places where we use it unconditionally though :/
<hkaiser> jbjnr: and no, I don't know brjsp
<jbjnr> ta
<jbjnr> "if constexpr" - are you sure about HPX_CONSTEXPR
<jbjnr> I'm just sticking to enable_if for now
<hkaiser> jbjnr: ahh 'constexpr if' - no that's out of scope for now
<hkaiser> sorry, misunderstood
<K-ballo> will take a couple years...
<jbjnr> yers :(
<jbjnr> ^years I mean - = sadness
jakemp has quit [Ping timeout: 240 seconds]
<parsa[w]> hkaiser: dot product updated
<hkaiser> k
<hkaiser> was it still broken?
<hkaiser> I thought that was ok already
patg[[w]] has joined #ste||ar
<hkaiser> parsa[w]: ahh you readded the 1dx2d case, thanks
<hkaiser> patg[[w]]: we'll leave 5:30 from here
parsa has joined #ste||ar
<parsa[w]> hkaiser: /phylanx/tests/unit/ast/to_string.cpp:45:9: error: no matching constructor for initialization of 'phylanx::ir::node_data<double>'
<parsa[w]> oh
<parsa[w]> that's obvious
eschnett has quit [Quit: eschnett]
<hkaiser> parsa[w]: I made the node_data constructors explicit
<parsa[w]> no that one's a rowVector issue
<hkaiser> k
<patg[[w]]> hkaiser: ok see you there
daissgr has quit [Remote host closed the connection]
parsa has quit [Ping timeout: 250 seconds]
patg[[w]] has quit [Quit: Leaving]
parsa has joined #ste||ar
<hkaiser> parsa: I'm through with the primitives, mul, div, sub are the only one missing, I believe
<parsa> div is done
<parsa> sub is done
<hkaiser> cool
<parsa> give me some time... i'll have all tests pass
<hkaiser> parsa: sure, absolutely
parsa has quit [Quit: Zzzzzzzzzzzz]
aserio has quit [Quit: aserio]
<jbjnr> hkaiser: my numa hint stuff is now working for all cases. hoping to get some first graphs out tomorrow. then I'll know if we have finally nailed the linear algebra project
<hkaiser> jbjnr: nice!
<jbjnr> I'm a little bit frightened ...
<jbjnr> if it doesn't speed things up ...
<hkaiser> jbjnr: I'm using your use case to force the committee into making the executor interface variadic
<jbjnr> which test case ? the async exec test I have been playig with ?
<hkaiser> no the case where you look at the arguments inside the executor
diehlpk_work has joined #ste||ar
<jbjnr> surely the interface is variadic alreDY ISN'T IT?
<jbjnr> oops
<hkaiser> no
eschnett has joined #ste||ar
<hkaiser> it's just us implementing it using a variadic interface
<jbjnr> ah.
<hkaiser> I couldn't convince them so far, but this is a strong use case
<jbjnr> if you need me to write anything down ....
<jbjnr> if this stuff works, then it's a very strong case.
<hkaiser> jbjnr: you could describe your use case here: https://github.com/executors/executors/issues/9
<jbjnr> I'm not sure I understand quite what you mean though. how could the executor not be variadic? invoke is variadic, async is too ... etc
<jbjnr> ^^
<hkaiser> right
eschnett has quit [Client Quit]
<hkaiser> currently the design just takes a function object, youhave tpo bind the args yourself - but that prevents the executor from looking at them
<jbjnr> ok. that makes more sense
<jbjnr> are you discussing this stuff this week?
<hkaiser> as you can see I'm asking for this for while now
<hkaiser> we discuss executors, but not this particular issue
<jbjnr> ok, then give me a few days and if I get good results, we can really use this as a driving example
<hkaiser> perfect
patg[[w]] has joined #ste||ar
EverYoung has joined #ste||ar
zbyerly_ has joined #ste||ar
EverYoun_ has quit [Ping timeout: 252 seconds]
parsa has joined #ste||ar
hkaiser has quit [Quit: bye]
patg[[w]] has quit [Quit: Leaving]
hkaiser has joined #ste||ar
<parsa[w]> hkaiser: generate_tree, define_operation, file_csv_primitives are broken
<hkaiser> parsa: ok, will fix
<hkaiser> the first two I know (I think those are already fixed on master), csv I will look
<parsa[w]> is there anything else i need to do on this branch?
<hkaiser> parsa[w]: no thanks - you should call it a day
<parsa[w]> when do you think these are going to be merged? they'll merge conflict my performance counter code
<parsa[w]> with*
<hkaiser> once the tests are green, I'll try to finish working on it tonight
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar