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 | Everybody: please respond to the documentation survey: https://forms.gle/aCpNkhjSfW55isXGA
tufei has joined #ste||ar
hkaiser has quit [Quit: Bye!]
diehlpk_work has quit [Remote host closed the connection]
<rachitt_shah[m]>
I needed a bit of insight on this area - would it make sense to club together modules+API for a particular function into one page?
<rachitt_shah[m]>
We've already attached the API doc for the module in the module page, but just trying to understand if doing this might help?
<ms[m]>
rachitt_shah: I don't think I understand the question, could you rephrase? What exactly are you proposing?
<gnikunj[m]>
rachitt_shah I don't have any travel plans today. What time would you like to meet?
<rachitt_shah[m]>
gnikunj I'll ping you in a while? Probably around 4:00/5:00 PM IST works?
<gnikunj[m]>
sure, works for me
kxkamil has joined #ste||ar
<rachitt_shah[m]>
Sure, I'll send an invite for 5:00 PM
tufei_ has joined #ste||ar
tufei has quit [Remote host closed the connection]
<rachitt_shah[m]>
<ms[m] "rachitt_shah: I don't think I un"> Would it be alright if I share it on the call?
<gnikunj[m]>
<ms[m] "rachitt_shah: I don't think I un"> ms: I think rachitt_shah wants to merge modules with API reference such that API reference opens up to module classes (instead of current Public and Full API list). Each class (Core, parallelism etc.) will have submodules and functions clicking on which will lead to the API side of things.
<rachitt_shah[m]>
Also, sending a PR in prior to the meeting
<rachitt_shah[m]>
Basic changes, then we could jump into niche areas if the basics look good enough
hkaiser has joined #ste||ar
<rachitt_shah[m]>
ms: could you help me with doxygen a bit? I've had a config file generated inside my hpx folder for doxygen, but when I run "doxygen [configfile]" the docs don't build
<rori[m]>
have you checked the permissions of those files?
<rachitt_shah[m]>
Yep, I ran as sudo, I was able to get an output for arch.cpp, so I'm guessing my config file needs work
<rori[m]>
I don't know if you are on linux, if yes I think a simple chmod -R would do
<rachitt_shah[m]>
My file patterns for the config seems to be good as well
<rachitt_shah[m]>
Changed them, yet similar issues
K-ballo has joined #ste||ar
<ms[m]>
rachitt_shah: what are the input directories relative to? Where is your doxygenfile? Do those directories exist? The paths look like they're for outside docker but you might be running things inside docker?
<rachitt_shah[m]>
Doxygenfile is in the folder I'm running doxyen in, the directories do exist, and I'm build it outside docker, got a large number of errors inside docker so ran it completely outside
<ms[m]>
If you're running outside docker why are the input directories inside your docker build directory? E.g. /home/rachitt/Desktop/hpx/build/docker-hpx/docs/hpx
<rachitt_shah[m]>
I'll try to run it in the hpx directory then
<rachitt_shah[m]>
That might work?
<ms[m]>
You'd likely want to point things to the hpx source directory, not the build directory
<ms[m]>
If the hpx directory means the root source directory, yes
<rachitt_shah[m]>
Got it, running from source
<rachitt_shah[m]>
Nope, got the same when running from source, would you like to see my Doxyfile?
<rachitt_shah[m]>
The only major changes I've made are to toggle INLINE_SOURCES = YES to show the file from which it's building
<ms[m]>
I can't help you much more right now, but yes, please share it
<rachitt_shah[m]>
I did, I have been following this, I would run through it again to make sure I'm not missing anything
<rachitt_shah[m]>
Let me try it again, sorry about the messing up(yet again)
hkaiser has joined #ste||ar
<rachitt_shah[m]>
Also, remember we last spoke about printing files from which docs are generated from? I got a way to do it with sphinx, didn't try yet, but if we add .. literalinclude:: filename in conf.py.in, it might just work
<rachitt_shah[m]>
Would it work if I temporarily remove conf.py.in and just run doxygen?
<gnikunj[m]>
hkaiser: do we have an option to invoke a stackless HPX thread?
<hkaiser>
gnikunj[m]: yes
<gnikunj[m]>
hkaiser: Nice, how do we do that? (Also, GSoD meeting)
tufei_ has quit [Quit: Leaving]
tufei has joined #ste||ar
<gnikunj[m]>
hkaiser: do that with futures ^^
<ms[m]>
gnikunj: parallel executor with nostack stacksize
<gnikunj[m]>
ok. Any gotchas I need to remember while using nostack option?
<ms[m]>
Have a look at with_stacksize, you should be able to apply that to the parallel executor
<gnikunj[m]>
Ah ok. Let me have a look.
<ms[m]>
No yielding or suspending obviously (or you'll yield or suspend the OS thread) but otherwise nothing we know of, but then we haven't used them much
<gnikunj[m]>
nice!
<jedi18[m]>
hkaiser: K-ballo Any idea which compiler would support ranges::iter_swap?
<jedi18[m]>
clang12 dosen't support it
<jedi18[m]>
gcc dosen't have it listed in the cxx status page
<K-ballo>
any standard library that supports ranges should provide it