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/
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 256 seconds]
diehlpk has quit [Remote host closed the connection]
diehlpk has joined #ste||ar
EverYoun_ has quit [Read error: Connection reset by peer]
EverYoung has joined #ste||ar
vamatya has quit [Ping timeout: 256 seconds]
EverYoung has quit [Ping timeout: 255 seconds]
apsknight has quit [Ping timeout: 260 seconds]
vamatya has joined #ste||ar
apsknight has joined #ste||ar
hkaiser has quit [Quit: bye]
katywilliams has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
parsa has quit [Read error: Connection reset by peer]
parsa has joined #ste||ar
jaafar has quit [Ping timeout: 260 seconds]
katywilliams has quit [Ping timeout: 276 seconds]
diehlpk has quit [Ping timeout: 260 seconds]
diehlpk_ has quit [Ping timeout: 255 seconds]
katywilliams has joined #ste||ar
katywilliams has quit [Ping timeout: 240 seconds]
K-ballo has quit [Quit: K-ballo]
apsknight1 has joined #ste||ar
anushi has joined #ste||ar
apsknight has quit [Ping timeout: 256 seconds]
apsknight1 is now known as apsknight
vamatya has quit [Ping timeout: 264 seconds]
vamatya has joined #ste||ar
EverYoung has joined #ste||ar
nanashi55 has quit [Ping timeout: 264 seconds]
nanashi55 has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
katywilliams has joined #ste||ar
katywilliams has quit [Ping timeout: 256 seconds]
Anushi1998 has joined #ste||ar
jaafar has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
jaafar has quit [Ping timeout: 248 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Client Quit]
Anushi1998 has quit [Remote host closed the connection]
<hkaiser>
we're using hpx::util::format in Phylanx - do I need to change the format strings now?
<K-ballo>
yes
<hkaiser>
replace %1%?
<K-ballo>
%1% with {1}, %xwy with {:xwy} is the simplest minimal change
<K-ballo>
except for absolute tabulations, those are not supported anymore
<hkaiser>
is there a way to know which format to use (if I don't know which hpx version I have)?
<K-ballo>
no
Anushi1998 has quit [Quit: Leaving]
<hkaiser>
uhh
<K-ballo>
<= 1.1 old format, else new format
<hkaiser>
ok, makes sense
apsknight1 has joined #ste||ar
apsknight has quit [Ping timeout: 248 seconds]
apsknight1 is now known as apsknight
<nikunj_>
@hkaiser: I have good news! We could simply avoid global initialization. I have found a very simple implementation of the 2nd method I wrote in my proposal (the one which had overriding compiler functions)
diehlpk has joined #ste||ar
diehlpk_ has joined #ste||ar
<nikunj_>
@hkaiser: I wrote there that I will have to override each one of them, turns out doing only 1 would actually suffice. It will involve overriding only __libc_start_main
parsa has quit [Quit: Zzzzzzzzzzzz]
<nikunj_>
@hkaiser: The way I found uses shared library concept. I simply preload my shared library with redefinition of __libc_start_main. This time it does not call "main" but a "custom_main" function.
<nikunj_>
@hkaiser: This would mean that if you could call hpx::init from "custom_main" to "main", main will be called on an hpx thread
<nikunj_>
@hkaiser: This method will simply not have any initialization sequence problems since we are changing the entry point from main to our entry point custom_main
<hkaiser>
nikunj_: sounds good
<hkaiser>
nikunj_: this will work for libstdc++ (gcc) only, right?
<nikunj_>
@hkaiser: at the moment i can only confirm for libstdc++ (gcc) only
<nikunj_>
I haven't tried it with other compilers
<hkaiser>
nod, as expected this is very platform specific
<hkaiser>
let's do one step at a time
<nikunj_>
@hkaiser: yes, I can search how to make things work for clang
<hkaiser>
(and perhaps msvc)
<nikunj_>
Also I do have one prototype if you'd like to try out yourself (it does not have hpx stuff but I could add it later when GSoC period starts)
<hkaiser>
k
<hkaiser>
nikunj_: you still need to do all the usual initialization in that routine
<hkaiser>
all global constructors need to be called, etc.
<K-ballo>
I don't even have the connectors to access media from that time :P
<nikunj_>
@hkaiser: I don't think there will be need to call the global constructor
<nikunj_>
I am just replacing main as entry point that's it
<K-ballo>
no seriously, I don't think I have access to IDE connectors anymore
<nikunj_>
it's more like doing #define main, but elegantly
<hkaiser>
sure, had to ask
<hkaiser>
nanashi55: nice
<hkaiser>
so global constructors will still be called by the runtime
<hkaiser>
nikunj_: ^^
<nikunj_>
@hkaiser: exactly
<nikunj_>
its more like hooking into __libc_start_main and calling custom_main instead of actual main
<nikunj_>
This way we just changed the entry point from main to custom main
<nikunj_>
everything else still remains the saim
<nikunj_>
*same
<hkaiser>
nice
<nikunj_>
I could search for ways to implement the same for clang and msvc as well
<nikunj_>
This would simply things a lot
<hkaiser>
yes, please
<hkaiser>
nikunj_: this should end up being very similar to our #define main solution (implementation-wise)
<nikunj_>
@hkaiser: yes it should, but this way won't bring any unexpected results
diehlpk_ has quit [Ping timeout: 256 seconds]
<hkaiser>
right
<nikunj_>
@hkaiser: this way is as robust as calling hpx::init() from main
diehlpk has quit [Ping timeout: 276 seconds]
<hkaiser>
nod
<nikunj_>
@hkaiser: In my proposal I wrote I'll be implementing the global initialization way, but now I'll be implementing the overriding methods. Does it complicate things?
<hkaiser>
no
<nikunj_>
@hkaiser: phew, glad to know that :)
<hkaiser>
nikunj_: has your timings changed for the summer now?
<hkaiser>
have*
<hkaiser>
we will not be able to select your proposal if you plan to be not available for half of the time
<nikunj_>
@hkaiser: yes I changed them, this is very interesting and I can easily devote more time
<nikunj_>
I changed the timings from 35-45h a week to 40-50h post my college reopens
<nikunj_>
This way I will definitely devote atleast 40h after my college reopens
<hkaiser>
no need to plan for more than 40h/week, but it needs to be consistent over the whole GSoC period
<nikunj_>
during my holidays, I can easily devote 60-70h a week
<nikunj_>
and I have 2months holidays during the 3month duration
<nikunj_>
So this would mean that I can work 60-70h a week for 2months and then 40-50h a week the last month
<hkaiser>
k
<zao>
Going above fulltime? Sounds silly.
<hkaiser>
I agree
<hkaiser>
time will tell
<K-ballo>
developers and time estimates...
<nikunj_>
zao: what I'm doing links with information security, another field i'm interested in
<nikunj_>
the way I used in my example above was a way I used while solving one of the CTF question (CTF: Capture The Flag)
<nikunj_>
CTFs are basically hacking competitions where the a backdoor entry is deliberately open for exploitation
<nikunj_>
I just didn't realise it earlier that I could make use of this way to dynamically link my own shared library to make things work for hpx
<zao>
Does anyone of your sites run WLCG grid by the way?
<hkaiser>
I don't even know what this is
<zao>
Computations for the LHC experiments (CMS, ATLAS, ALICE) run in a distributed fashion out on individual sites that have pledged compute power.
<zao>
Weirdo mix of batch computing and cloud, way before clouds existed.
<zao>
We're part of the nordic T1 and have both storage and compute.
Anushi1998 has joined #ste||ar
diehlpk has joined #ste||ar
diehlpk_ has joined #ste||ar
parsa[w] has quit [Read error: Connection reset by peer]
parsa[w] has joined #ste||ar
parsa[w] has quit [Read error: Connection reset by peer]
EverYoung has joined #ste||ar
parsa[w] has joined #ste||ar
jaafar has joined #ste||ar
EverYoung has quit [Ping timeout: 256 seconds]
apsknight has quit [Quit: apsknight]
diehlpk has quit [Ping timeout: 256 seconds]
diehlpk_ has quit [Ping timeout: 260 seconds]
verganz has joined #ste||ar
nikunj_ has quit [Quit: Page closed]
Anushi1998 has quit [Quit: Leaving]
quaz0r has quit [Ping timeout: 268 seconds]
quaz0r has joined #ste||ar
jaafar has quit [Ping timeout: 260 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 256 seconds]
anushi has quit [Ping timeout: 255 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 256 seconds]
anushi has joined #ste||ar
apsknight has joined #ste||ar
parsa[w] has quit [Read error: Connection reset by peer]