hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar-group.org | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | This channel is logged: irclog.cct.lsu.edu
K-ballo has quit [Quit: K-ballo]
hkaiser has quit [Quit: Bye!]
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
diehlpk_work has joined #ste||ar
<dkaratza[m]>
hkaiser: sorry i missed that it was earlier
<dkaratza[m]>
i can join now if you want
<hkaiser>
dkaratza[m]: let's move it to next week, sorry for the misunderstanding
<dkaratza[m]>
hkaiser: I'm so sorry, we can do it tomorrow as well or later today so that we don't have to skip. But whatever you prefer
<gonidelis[m]>
thank god daylight savings were banned yesterday
<hkaiser>
who banned it?
<gonidelis[m]>
the senate, no?
<gonidelis[m]>
well, not "banned"
<gonidelis[m]>
i am just being dramatic
<gonidelis[m]>
hkaiser: can we meet now?
<hkaiser>
gonidelis[m]: sure, sec
<K-ballo>
what's "banned" dramatic for?
<gonidelis[m]>
it's not actually banned
<gonidelis[m]>
quite the opposite
<gonidelis[m]>
it's permanent
<gonidelis[m]>
only "ban" is a strong word
<hkaiser>
gonidelis[m]: I'm ready to meet
<gonidelis[m]>
hkaiser: do you think that the cmake flags removed are a breaking change?
<hkaiser>
let's list them as such
<gonidelis[m]>
hkaiser: K-ballo is tag_invoke just a way to hide function names from the global scope? i feel like hiding the names is good, but i don't know if the reason is correct. for me is just like grouping and tidying up your socks in the dresser instead of having to look for and pick them up from the pile in your bed when you are getting prepared for the night out
<gonidelis[m]>
kinda like speeding up the lookup (searching?)
<gonidelis[m]>
but is that right?
<K-ballo>
not really
<K-ballo>
have you read the corresponding paper?
<gonidelis[m]>
many times have started reading it. never made it to the end.
<K-ballo>
did you get thru the motivations section?
<gonidelis[m]>
and yeah the user can reuse the same name for their own purposes
<gonidelis[m]>
yes
<K-ballo>
it's a generalized customization-point object mechanism
<K-ballo>
a CPO "factory" if you want
<gonidelis[m]>
a thing that creates a thing that does overloading
<K-ballo>
the difference between hand rolling your own CPO or using tag_invoke is just boilerplate
<K-ballo>
any other advantages come from CPOs in general, not tag_invoke in particular
<gonidelis[m]>
alright
<gonidelis[m]>
facilitation
<gonidelis[m]>
i think i was having some difficulties distinguishing CPOs in general and tag_invoke
<gonidelis[m]>
mainly becuase the only cpo that i know of is tag_invoke
<K-ballo>
you have read the CPO design paper, have you not? I'm thinking of somebody else?
<K-ballo>
ranges::begin/end where some of the first CPOs, along with swap
<gonidelis[m]>
somebody else?
<gonidelis[m]>
ah
<gonidelis[m]>
i read "you have to"
<gonidelis[m]>
yes i have posted questions on that paper before. i haven't managed to finish it. will give it another shot
<gonidelis[m]>
the object was the algorithm that could be used either with .begin .end or with the (custmoized) rng arg?
<gonidelis[m]>
(yor begin, sent args
<gonidelis[m]>
)
<K-ballo>
uhm?
<gonidelis[m]>
why is ranges::begin/end a cpo?
<K-ballo>
a CP because there exist ranges other than those defined by ranges:: itself
<K-ballo>
an O because go read the CPO design paper
<gonidelis[m]>
lol
<gonidelis[m]>
thanks
<dimpapag[m]>
https://github.com/z3y50n/pyhpx Here is a repo I created for playing with pybind11 and hpx. I have added a wrapper around hpx's is_heap, some benchmarks and did some initial reading on the building system
rtohid[m] has joined #ste||ar
<hkaiser>
dimpapag[m]: nice
<hkaiser>
dimpapag[m]: if you're interested in language bindings, this might be a good resource as well: https://github.com/Tolc-Software/
<gonidelis[m]>
dimpapag: !!! wow
diehlpk_work has quit [Remote host closed the connection]
<dkaratza[m]>
hkaiser: I updated the PR, also I fixed the full API pages issue (also included in the PR). So now you can merge that and if you have any concerns arising from the changes I made, we can discuss in the next meeting