K-ballo 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/
<gonidelis[m]> i need to revise our dialogue actually
<K-ballo> yes
<K-ballo> some form of B_view{C}
<gonidelis[m]> K-ballo: could it actually be that eric's `indirect_relation` is different from c++20's `indirectly_comparable`?
<K-ballo> you know who'd know the answer to that question?
<gonidelis[m]> a guy who lives in a country where we have an 8 hour difference. the only reeason i keep asking you is because it's 3am here and I know i couldn't sleep without finding the incosistency
<K-ballo> 5pm is a very decent time to be answering random questions
<K-ballo> 9pm not so much
<gonidelis[m]> You are 9pm?
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
K-ballo has quit [Quit: K-ballo]
<jbjnr> I'm not able to log into the google summer of code site - is that because I'm not a mentor this year, or is the site down/flakey
<ms[m]1> jbjnr: works for me (tm)
<ms[m]1> did you register at some point?
<jbjnr> I don't think I got an invite this year, so probably no.
<ms[m]1> hrm, ask patrick to add you... I think it should still be possible to add you
<ms[m]1> itn: hey, sorry for not replying earlier to your questions and thanks for opening your PR
<ms[m]1> given that the deadline for proposals is coming up soon I'd recommend you work on your proposal for the rest of the time
<ms[m]1> if you have time to keep working on the pr that's nice as well, but prioritise the proposal
<ms[m]1> as a general tip for the pr, try to get it into a shape where you can run tests on your newly added algorithms successfully, we can then take it from there to suggest improvements to style, performance, etc.
K-ballo has joined #ste||ar
<gonidelis[m]> K-ballo what's this online cpp debugger you are using?
<gonidelis[m]> ah.... explorer
<K-ballo> compiler-explorer isn't a debugger
<gonidelis[m]> yeah sorry
<gonidelis[m]> assembly converter
<K-ballo> I don't think I've heard of any C++ online debugger
<gonidelis[m]> idk if it's worth though
<ms[m]1> gonidelis[m]: what does this comment mean: https://github.com/STEllAR-GROUP/hpx/pull/5225#discussion_r609476951?
<gonidelis[m]> ms[m]: ah sorry
<gonidelis[m]> i meant I aggree with your comment and will do what you suggested
<ms[m]1> all right :P
<gonidelis[m]> shouldn't use a militaristic term
hkaiser has joined #ste||ar
<hkaiser> ms[m]1: yt?
<ms[m]1> hkaiser: here
<hkaiser> see pm, pls
<srinivasyadav227> hkaiser: could you explain this one https://github.com/STEllAR-GROUP/hpx/blob/master/libs/parallelism/algorithms/include/hpx/parallel/datapar/loop.hpp#L248?, what does is_data_aligned() function do?
<srinivasyadav227> does it return true when Iter is aligned?
<itn[m]> <ms[m]1 "given that the deadline for prop"> Thanks ms for reviewing and guiding me with some important tips about my work, Now as suggested by you, I'll be more focused on preparing my proposal.
<hkaiser> it checks whether the elements in the input range are aligned such vectorization could be done directly
<srinivasyadav227> so it returns true when it is aligned?
<hkaiser> itn[m]: thanks for starting to work on the shift algorithms
<hkaiser> what would you like to propose for GSoC?
<hkaiser> srinivasyadav227: yes
<srinivasyadav227> hkaiser: then shouldnt it be " !detail::is_data_aligned " ?, the prologue loop should run until the iterator is aligned and len != 0,
<hkaiser> srinivasyadav227: nice catch!
<hkaiser> srinivasyadav227: aparently the same is wrong in other places as well
<srinivasyadav227> cool, thanks, i will change it soon
<srinivasyadav227> yes,
<hkaiser> srinivasyadav227: thanks!
<srinivasyadav227> and count was variable was used instead of len, in vector calls i.e callv(...), i fixed this
<hkaiser> srinivasyadav227: I'm now suprised anything worked at all ;-)
<hkaiser> please add regression tests that verify things are ok now
<gnikunj[m]> hkaiser: lol when was all this implemented?
<gnikunj[m]> also, were you short on coffee then ;-)
<hkaiser> 2014?
<hkaiser> yah, most likely
<gnikunj[m]> damn that's way back in time
<gnikunj[m]> srinivasyadav227: thanks a lot for looking into it!
<srinivasyadav227> hkaiser: yes, the unit tests didnt help much, when i checked with other inputs it failed ( it passed with 10007 default test luckily)
<hkaiser> thanks
<srinivasyadav227> gnikunj[m]: thank you ;-)
hkaiser has quit [Quit: bye]
<gnikunj[m]> ms: don't we have the hpx meeting today?
<gnikunj[m]> it doesn't show in my calendar
<ms[m]1> gnikunj: no, it was moved to next week (hartmut is away this week and next)
<gnikunj[m]> aah, got it. Thanks!
<gonidelis[m]> any ideas how to include range-v3 on wandbox?
<gonidelis[m]> any idea on that output https://wandbox.org/permlink/vw8HGxahb8oeGxuF ?
<K-ballo> you'd have to copy each and every range-v3 file you want to include
<K-ballo> it's all already there? that's convenient
<K-ballo> the usual trick to spell out types is trying to instantiate an incomplete template
<gonidelis[m]> actually it worked... i don't why but it did
<gonidelis[m]> yeah it was there for some reason
<gonidelis[m]> K-ballo: hm... what do you mean about the incomplete template?
<K-ballo> incomplete means declared but not defined
<gonidelis[m]> if i instantiate the type won't it be complete then?
<gonidelis[m]> the template*
<K-ballo> how?
<gonidelis[m]> i pass a specific type to it
<gonidelis[m]> no?
<K-ballo> how would instantiating a template produce a definition?
<gonidelis[m]> right
<gonidelis[m]> ok
<gonidelis[m]> should i create an incomplete function template or class tempolate?
<gonidelis[m]> and you mean instantiate it with decltype(ax>
<gonidelis[m]> `decltype(ax)` *
<gonidelis[m]> ?
<gonidelis[m]> K-ballo: same output
<K-ballo> any way in which an instantiation would be viable will do
<K-ballo> typetest is not an incomplete type, it has a definition
<gonidelis[m]> well if i remove the `{}`s it won't compile
<gonidelis[m]> so i guess passing decltype(ax) is not viable
<gonidelis[m]> wait...
<gonidelis[m]> i get the type in the error message though
<gonidelis[m]> is that the catch?
<K-ballo> precisely
<gonidelis[m]> only C++ would have errors that hack C++
<K-ballo> the very first template metaprogram computed prime numbers and listed them in compiler errors
<K-ballo> it's the only "out" the compiler has for humans
<gonidelis[m]> K-ballo: yes. i remember from hartmut's lectures
<gonidelis[m]> ms[m]: have we adapted any algo from numeric at all?
<K-ballo> "adapted" as in?
<gonidelis[m]> K-ballo: ranges overloads
<K-ballo> there are no ranges overloads in numeric
<gonidelis[m]> K-ballo: yeah just saw that
<gonidelis[m]> why though?
<gonidelis[m]> i think niebler provides them
<K-ballo> we've been through this a couple times already
<K-ballo> coming up with the numeric concepts was too hard
<K-ballo> was left as a todo
<gonidelis[m]> we as the committee?
<K-ballo> we as you and I and the rest of this channel
<K-ballo> we've discussed the lack of range-ification of <numerics> at least thrice already
<gonidelis[m]> lol
<gonidelis[m]> i am very scared that you remember that
<K-ballo> I'm surprised you don't, starting to imagine it's actually several students under one nick/persona
<K-ballo> (not serious)
<K-ballo> on a serious note though, make sure you sleep well, it's important
<gonidelis[m]> i do now that you are mentioning. it's just that i am exposed to these concepts for the first time so I might didn't even know what numeric actually is, when i was asking you for the first time.
<K-ballo> numeric is a header
<gonidelis[m]> yeah i know it now
<gonidelis[m]> but the first time i asked sth relevant you were like "this is numeric, not algorithm" and i was like what's the dff, so i wouldn't expect my past self to remember that is not even range-ified
<K-ballo> "hmm, and there is no ranges::adjacent_difference", gonidelis[m]: "because <numeric> ?"
<gonidelis[m]> sleep? i need to parallelize ranges until the end of april or else you won't be the only one giving me a hard time anymore, probably most of the cppnow attendees will too
<gonidelis[m]> (i appreciate your giving though)
<gonidelis[m]> <K-ballo ""hmm, and there is no ranges::ad"> i don't recall what i was trying to say but i don't claim denying its existance
<gonidelis[m]> first time i heard was late august i think