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
parsa has quit [Read error: Connection reset by peer]
parsa| has joined #ste||ar
<github>
[hpx] sithhell closed pull request #3306: Marking migratable objects in their gid to allow not handling migration in AGAS (master...mark_migratable_objects) https://git.io/vpE2c
<github>
[hpx] sithhell deleted mark_migratable_objects at 83c0f3e: https://git.io/vpa6U
<github>
[hpx] sithhell opened pull request #3309: Fixing potential stack overflow for dataflow (master...fix_dataflow_stack_overflow) https://git.io/vpa6T
<github>
[hpx] hkaiser created fixing_phylanx_395 (+1 new commit): https://git.io/vpVlH
<github>
hpx/fixing_phylanx_395 e0d74af Hartmut Kaiser: Don't overwrite user settings with defaults in registration database
<github>
[hpx] hkaiser opened pull request #3310: Don't overwrite user settings with defaults in registration database (master...fixing_phylanx_395) https://git.io/vpVlh
ASamir has joined #ste||ar
akheir has joined #ste||ar
jakub_golinowski has joined #ste||ar
Samir has joined #ste||ar
ASamir has quit [Quit: Page closed]
Samir is now known as Guest90696
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
bibek has joined #ste||ar
jakub_golinowski has quit [Ping timeout: 240 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
parsa has joined #ste||ar
Samir has joined #ste||ar
Samir is now known as Guest90065
Guest90696 has quit [Ping timeout: 260 seconds]
<github>
[hpx] K-ballo created format-escape (+1 new commit): https://git.io/vpV2P
<github>
hpx/format-escape 7e93999 Agustin K-ballo Berge: Fix escaping of closing braces in format specification syntax
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
Guest90065 has quit [Quit: Leaving]
<hkaiser>
K-ballo: P1017 is almost ready to be submitted, not sure if you'd like to give it a last read over
EverYoung has joined #ste||ar
<K-ballo>
sure, I'll make another pass
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
<hkaiser>
thanks
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 260 seconds]
nikunj has joined #ste||ar
<nikunj>
@hkaiser: I wanted to write a blog for what the work I've done and the share the overall experience. What platform should I use to write a blog?
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
EverYoung has joined #ste||ar
EverYoun_ has quit [Ping timeout: 255 seconds]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
<K-ballo>
hkaiser: there's inconsistent use of std:: qualifications, most functions are unqualified but a few here and there are
<github>
[hpx] K-ballo force-pushed format-escape from 7e93999 to afc299c: https://git.io/vpV9T
<github>
hpx/format-escape afc299c Agustin K-ballo Berge: Fix escaping of closing braces in format specification syntax
<github>
[hpx] K-ballo opened pull request #3311: Fix escaping of closing braces in format specification syntax (master...format-escape) https://git.io/vpV9s
aserio has quit [Ping timeout: 276 seconds]
aserio has joined #ste||ar
<aserio>
hkaiser: yt?
<hkaiser>
aserio: here
<aserio>
hkaiser: I think I will try to schedule the primitive talk Monday
<parsa[[w]]>
hkaiser: bad news. building inspect triggers build all again :))))
<aserio>
We already have 2 meetings that day anyway
<hkaiser>
ASamir: heller is most likely asleep right now
<ASamir>
Ok. Thanks.
<hkaiser>
ASamir: he's in Germany, FWIW
<heller>
Please excuse my ignorance, but what's fflib?
<hkaiser>
ahh, there he is ;)
<jbjnr>
ASamir: did you get my email with the comments from salvatore about the new library they will develop?
<hkaiser>
ASamir: ^^
<jbjnr>
it bounced first time
<jbjnr>
heller: fflib is the lib from ethz that implements collectives
<jbjnr>
the one we want to use for our gsoc stuff
<ASamir>
heller: A model proposed for collective communications
<heller>
Ahh, thanks
<jbjnr>
we = ASamir and I and with your help if ASamir needs extra parcelport assistnce when I'm not around
<heller>
Alright
<jbjnr>
they are creating a new fflib without portals
<heller>
Perfect
<jbjnr>
but using ibverbs instead
<heller>
Oh
<heller>
Why not libfabric?
<ASamir>
Yes John I've received the mail.
<jbjnr>
which is great for me\we can easily make it work with libfabric
<jbjnr>
great^
<jbjnr>
not libfabric because they don'r know about it and ibverbs is slightly lower level
<hkaiser>
will there be a fallback implementation for non-IB systems?
<ASamir>
I've made an account on CSCS. Is there any step to start with?
<jbjnr>
yes
<heller>
And as hkaiser said, I'm on the run to sleep ;)
<jbjnr>
ASamir: once you get into ela, then you can ssh into daint - that is the cray - you can start by experimenting with building stuff there
<jbjnr>
like hpx/fflib/libfabric any of that, just to get the feel of things
<jbjnr>
on monday we could have a chat if you want
<hkaiser>
jbjnr: will there be a fallback implementation for non-IB systems?
parsa has quit [Quit: Zzzzzzzzzzzz]
<ASamir>
Ok I will work on building libfabric until monday
<jbjnr>
hkaiser: they will provide a fallback using mpi itself (irony)
<hkaiser>
k
<hkaiser>
good
<jbjnr>
but we don't care, we will use libfabric and I have started to get the libfabric layer working on tcp
<hkaiser>
I'd like to avoid depending on IB with any higher level feature in HPX
<jbjnr>
so there's that
<hkaiser>
ahh, ok
<hkaiser>
jbjnr: so we can get rid of the asio based tcp pp as well?
<jbjnr>
in my dreams we could replace the tcp parcelport with the libfabric one, but that might not be possible because of bootstrapping issues
<jbjnr>
^^
<jbjnr>
crossed messages
<hkaiser>
ok
<jbjnr>
not 100% certain yet
<jbjnr>
and libfabri on tcp doesn't actually work yet
<jbjnr>
(endpoint creation isssues)
<jbjnr>
ASamir: what would be useful to get started is have a look at the fflib stuff and also the new repo they've started. what did you want to ask about fflib?
<jbjnr>
aha that link is to fflib2 good
<ASamir>
I didn't know where to start. The code does not have any documentation
<ASamir>
How should I tackle the code?
<jbjnr>
yup. just loking through it, it looks very basic, nothing yet implemented apart from some stubs
<jbjnr>
for now, just see if you can work out what it can/can't do. we should create a cmake file to build it
<jbjnr>
since no makefile is present
EverYoun_ has quit [Remote host closed the connection]
<jbjnr>
just to see if it compiles. I'll ask salvatore on monday about it
<heller>
jbjnr: you have my support for kicking all other parcelports in favor of libfabric
<jbjnr>
great. (not sure we'll manage it, but worth trying)
EverYoung has joined #ste||ar
<heller>
yeah
<heller>
would be good...
<heller>
if TCP works, we're fine
hkaiser has quit [Ping timeout: 276 seconds]
eschnett has quit [Quit: eschnett]
aserio has quit [Quit: aserio]
hkaiser has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
parsa has joined #ste||ar
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
EverYoun_ has quit [Ping timeout: 276 seconds]
EverYoung has joined #ste||ar
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
Samir has joined #ste||ar
ASamir has quit [Read error: Connection reset by peer]