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
diehlpk_work has quit [Remote host closed the connection]
<gonidelis[m]>
2hk
diehlpk has joined #ste||ar
<gonidelis[m]>
hkaiser: i reckon this is not a problem right ? `vtune: Warning: Hardware collection of CPU events is not possible on this system. Microarchitecture performance insights will not be available.
<gonidelis[m]>
`
<gonidelis[m]>
given that we do not dig into the architectural details
<hkaiser>
gonidelis[m]: right
<hkaiser>
you need root access for this
diehlpk has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
<gonidelis[m]>
hkaiser: see pm please
K-ballo has quit [Quit: K-ballo]
diehlpk has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
hkaiser has quit [Quit: Bye!]
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
K-ballo has joined #ste||ar
<gonidelis[m]>
hkaiser: did you see the standards ranges proposal email from yesterday
<hkaiser>
gonidelis[m]: yes
<gonidelis[m]>
hkaiser: any opinions?
<hkaiser>
gonidelis[m]: well, we have it, haven't we?
<gonidelis[m]>
ye wasn't referring to that so much though
<gonidelis[m]>
he put an InputIterator in his proposed specification
<gonidelis[m]>
that's what troubled me
<hkaiser>
ahh, good catch, didn't see that
<gonidelis[m]>
and then he just called the algorithm with std::exec::par
<hkaiser>
that will not fly
<gonidelis[m]>
I actually took the liberty of responding
<gonidelis[m]>
and pointed him to our algos
<hkaiser>
ok
<gonidelis[m]>
cool
<hkaiser>
srinivasyadav227: yt?
<srinivasyadav227>
hkaiser: yes
<hkaiser>
srinivasyadav227: I need to skip our meeting this morning
<srinivasyadav227>
oh np :)
<hkaiser>
sorry
<srinivasyadav227>
ok np :)
<srinivasyadav227>
hkaiser: no, its alright :)
<hkaiser>
thanks
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
<gonidelis[m]>
hkaiser: could you please remind me the command that check whether my hpx build is release or debug
<gonidelis[m]>
?
<hkaiser>
--hpx:info
<gonidelis[m]>
thank you
diehlpk has quit [Quit: Leaving.]
<hkaiser>
gonidelis[m]: FWIW, there is --hpx::help as well ;)
<gonidelis[m]>
i just couldn't recollect if it was an hpx or a cmake option
<hkaiser>
two things tbd: inspect target name, and finalize the list of options that should be forwarded, etc.
<ms[m]>
right, I've changed the name on the hpx-local branch (though I see that I may not have pushed that; in any case, that's separate)
<hkaiser>
ms[m]: ok
<ms[m]>
is there anything you would need from me for the list options? do we agree on them?
<hkaiser>
ms[m]: the list itself is probably not that important as it can be changed at any point
<hkaiser>
I would like to ensure that users that rely on scripts to build/install things dont' run into version differences causing problems
<ms[m]>
indeed, but let's try to agree on the ones we have there at the moment at least
<hkaiser>
yes
<hkaiser>
in the first ste, ideally nothing should change, but we might want to warn users that things may change
<ms[m]>
yes
<hkaiser>
ms[m]: here is the rationale: I know the octotiger people have their scripts (non-spack) that explicitly specify options and I know that they go back and forth between HPX versions
<ms[m]>
if you're concerned about the hpx_error I had in there earlier that was mainly for making sure I'm setting the "correct" options in CI, your change to make it a warning is the right thing
<hkaiser>
ok
<hkaiser>
so let me have another look later today and then we can go ahead with this
<ms[m]>
I'm not at all opposed to the compatibility layer, I'd just like to finalize which options give a warning saying the local option instead and which ones are "duplicated" and the value is forwarded automatically to the local one
<hkaiser>
nod
<ms[m]>
*saying to use the local option instead
diehlpk has joined #ste||ar
diehlpk_work has joined #ste||ar
<hkaiser>
ms[m]: yt?
<gonidelis[m]>
hkaiser: do i needrelease type when building with apex? or relwithdeb /
<gonidelis[m]>
?
<hkaiser>
you need debugging information fo rapex to be able resolve addresses to names