hkaiser 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/ | GSoC2018: https://wp.me/p4pxJf-k1
hkaiser has quit [Quit: bye]
parsa has quit [Quit: Zzzzzzzzzzzz]
nanashi55 has quit [Ping timeout: 260 seconds]
nanashi55 has joined #ste||ar
verganz has joined #ste||ar
verganz has quit [Ping timeout: 260 seconds]
verganz has joined #ste||ar
verganz has quit [Client Quit]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 240 seconds]
simbergm has quit [Ping timeout: 268 seconds]
nikunj has joined #ste||ar
simbergm has joined #ste||ar
<github>
[hpx] msimberg created msimberg-patch-1 (+1 new commit): https://git.io/vp9bg
<github>
hpx/msimberg-patch-1 ab11219 Mikael Simberg: Fix CircleCI config to really push docker image only on master
<github>
[hpx] msimberg opened pull request #3315: Fix CircleCI config to really push docker image only on master (master...msimberg-patch-1) https://git.io/vp9bV
mcopik has joined #ste||ar
<jbjnr>
I've just noticed that on github I have 500 commits to HPX!
<heller>
yay
<heller>
i have a red pen
<jbjnr>
a red pen. Wow. that's cool. What will you use it for?
<jbjnr>
</sarcasm>
jakub_golinowski has joined #ste||ar
<heller>
jbjnr: marking my scars
<jbjnr>
self harming now? I suggest booze instead.
<heller>
yeah
<heller>
na, just going through my dissertation
<heller>
booze might help
<github>
[hpx] msimberg opened pull request #3316: Fix includes for 1d_stencil_*_omp examples (master...fix-1d_stencil-omp-examples) https://git.io/vpHvC
<jbjnr>
simbergm: yt?
<jbjnr>
if you are - you're right. HPX+APEX, is broken
<simbergm>
heller: circleci still did the push image step on my branch, so please don't merge
<jbjnr>
?
<simbergm>
sorry, unrelated
<simbergm>
circleci is trying to push docker images also for PRs when it should do it only for master
Viraj has joined #ste||ar
Viraj_ has joined #ste||ar
Viraj has quit [Ping timeout: 260 seconds]
<github>
[hpx] biddisco pushed 1 new commit to apex_task_wrapper_memory_bug: https://git.io/vpHqV
<github>
hpx/apex_task_wrapper_memory_bug 4915b07 John Biddiscombe: Fix a missing export keyword for apex_new_task and rebind_base_apex
verganz has quit [Quit: Leaving.]
verganz has joined #ste||ar
Viraj_ has quit [Ping timeout: 260 seconds]
Viraj has joined #ste||ar
verganz1 has joined #ste||ar
verganz has quit [Ping timeout: 265 seconds]
nikunj has quit [Ping timeout: 255 seconds]
hkaiser has joined #ste||ar
verganz1 has left #ste||ar [#ste||ar]
verganz1 has joined #ste||ar
verganz1 has left #ste||ar [#ste||ar]
verganz1 has joined #ste||ar
parsa has joined #ste||ar
Viraj has quit [Ping timeout: 260 seconds]
<hkaiser>
heller: yt?
<jbjnr>
hkaiser: you are not listed on the ISO c++ meeting doodle thingy. Are you still going?
<hkaiser>
I am
<jbjnr>
did you find a hotel for the week?
<hkaiser>
yes
<jbjnr>
can you remember where?
<hkaiser>
jbjnr: Noah Hotel in Ruetli
<jbjnr>
that seems far far away.
<hkaiser>
jbjnr: Rueti, that is
<hkaiser>
yah, I'll have a car
<jbjnr>
aha. better
<hkaiser>
jbjnr: I can give you a ride if you stay there as well
<jakub_golinowski>
simbergm, yt?
parsa has quit [Quit: Zzzzzzzzzzzz]
<simbergm>
jakub_golinowski: here
<jakub_golinowski>
Hello!
<simbergm>
how's it going?
<jakub_golinowski>
As I wrote in the e-mail the last week was mostly focused on pushing the ETH workload and I think it was quite successful. Yesterday, late evening, I managed to finish the project I referred to in the e-mail. So now I only have one project left that I will have to work on in parallel to GSoC which is managable.
<jakub_golinowski>
Today I started with walk-through of the resource-partitioner examples and wrote down a list of questions in the gdoc
<jakub_golinowski>
Could you have a look at them?
<simbergm>
sounds good, how much work do you think this other project will be per week?
<simbergm>
yep, I'll have a look
<simbergm>
just a minute
<simbergm>
hrm, is it the same document you linked to in the email? looks the same as last week to me
<jakub_golinowski>
I use the different gdoc
<jakub_golinowski>
wanted to add you by e-mail
<jakub_golinowski>
such that when you add comments you are not anonymous but it requires gmail e-mail
<simbergm>
ah, could you share it with mikael.simberg@iki.fi in that case
<jakub_golinowski>
ok, you should get an e-mail
<simbergm>
got it, thanks
<jakub_golinowski>
simbergm, the questions from today start at page 3
Viraj has joined #ste||ar
<simbergm>
ok, good, I'll answer inline in the doc so you have it written down and then we can discuss what doesn't make sense
<simbergm>
give me 15 mins or so
<jakub_golinowski>
thank you very much
<Viraj>
Hi everyone, I was building HPX with APEX and got some errors
<Viraj>
"error: no matching function for call to ‘new_task(const char*, long unsigned int, apex::task_wrapper*)’ UINTMAX_MAX, (apex::task_wrapper*)parent_task);"
<Viraj>
and so on..
<Viraj>
does anybody have any idea on how to fix?
<Viraj>
seems like it is not able to find the function defs?
<hkaiser>
Viraj: have you used recent versions of hpx and apex?
<Viraj>
yes i have
<Viraj>
do i have to change to specific branch?
mbremer has joined #ste||ar
<hkaiser>
Viraj: no idea - please send an email to 'Kevin Huck' <khuck@cs.uoregon.edu> and cc me
eschnett has quit [Quit: eschnett]
diehlpk has joined #ste||ar
Viraj_ has joined #ste||ar
Viraj has quit [Ping timeout: 260 seconds]
<jbjnr>
Viraj_: Make sure you update the apex subdir to latest master of the apex repo - then you might want to merge this branch into your hpx https://github.com/STEllAR-GROUP/hpx/pull/3298
aserio has joined #ste||ar
<jbjnr>
I pushed a fix to it a couple of hours ago as I had the same trouble
hkaiser has quit [Ping timeout: 265 seconds]
david_pfander1 has joined #ste||ar
<Viraj_>
okay so now i will update hpx.. also last time kevin told me that develop branch is more stable than master so I will replace the apex subdir with this branch of the apex repo. This should work right?
RostamLog has joined #ste||ar
mcopik has quit [Ping timeout: 268 seconds]
eschnett has joined #ste||ar
Viraj_ has quit [Ping timeout: 260 seconds]
mcopik has joined #ste||ar
hkaiser has joined #ste||ar
<heller>
hkaiser: Hey
<hkaiser>
heller: see pm, pls
Viraj has joined #ste||ar
Viraj has quit [Ping timeout: 260 seconds]
parsa[w] has quit [Read error: Connection reset by peer]
Viraj has joined #ste||ar
mcopik has quit [Ping timeout: 260 seconds]
parsa[w] has joined #ste||ar
mcopik has joined #ste||ar
aserio1 has joined #ste||ar
aserio has quit [Ping timeout: 260 seconds]
aserio1 is now known as aserio
nikunj has joined #ste||ar
<jakub_golinowski>
I have a question about oversubscription:
<jakub_golinowski>
In the default use-case the hpx is scheduling as many worker threads as #PUs on the given machine. So there is a single worker thread per core. However on top of that there are for example IO thread pool and Timers thread pool, with their threads. Does it mean oversubscription? In other words, is oversubscription (as in oversubscribing_resource_partitioner) defined as assigning more os-threads to core than one? Also are the t
<jakub_golinowski>
hreads from IO and Timer pools pinned to some cores?
simbergm has quit [Ping timeout: 268 seconds]
verganz1 has quit [Ping timeout: 248 seconds]
verganz has joined #ste||ar
aserio has quit [Ping timeout: 255 seconds]
Viraj_ has joined #ste||ar
Viraj has quit [Ping timeout: 260 seconds]
aserio has joined #ste||ar
aserio1 has joined #ste||ar
<nikunj>
how is hpx::util::function_nonser different than the std::function?
<K-ballo>
ideally it isn't, functionality-wise
aserio has quit [Ping timeout: 240 seconds]
aserio1 is now known as aserio
<nikunj>
K-ballo: I see
<nikunj>
K-ballo: was there a specific need for creating it instead of using the std::function?
<nikunj>
I'm just curious to know ;)
<K-ballo>
we didn't even had std::function available back then
<K-ballo>
the implementation conformance was all over the place
<nikunj>
K-ballo: oh, now I get it
<K-ballo>
and we had the machinery since we needed it for `hpx::util::function`
<nikunj>
nod
khuck has joined #ste||ar
khuck has quit [Remote host closed the connection]
simbergm_ has joined #ste||ar
Viraj_ has quit [Ping timeout: 260 seconds]
nikunj97 has joined #ste||ar
Viraj has joined #ste||ar
nikunj has quit [Ping timeout: 255 seconds]
<simbergm_>
jakub_golinowski: cooking dinner and then I'll get to your next question
<nikunj97>
the current hpx::init considers a function of type int(int, char**), however the main pointer (the pointer to the compiler's function) is of type int(int, char**, char**). Is there a way to cast it to int(int, char**) or would I have to implement the hpx::init from scratch for it?
<nikunj97>
*pointer to compiler's version of main ^^
hkaiser has quit [Ping timeout: 260 seconds]
verganz has joined #ste||ar
hkaiser has joined #ste||ar
<hkaiser>
nikunj97: provide your own main and call hpx_main from it
<nikunj97>
oh, I never thought of it. I could simply call hpx::init, and also define hpx_main in my file. I'll then call the main from there
<nikunj97>
this should prevent all the hassle
<nikunj97>
hkaiser: thanks
<hkaiser>
:D
<hkaiser>
nikunj97: you can solve all (but one) problem in computer science by adding yet another indirection layer
<nikunj97>
hkaiser: I am currently working on a better (more refined) version of yesterdays implementation. Should be able to reproduce it into code by the end of this day
<nikunj97>
hkaiser: yes you can!
<hkaiser>
nikunj97: nice
<nikunj97>
hkaiser: i will keep you updated, will share the link as soon i get it done
verganz has quit [Quit: Leaving.]
eschnett has quit [Quit: eschnett]
mcopik has joined #ste||ar
khuck has quit []
mcopik has quit [Read error: Connection reset by peer]