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/
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
vamatya has quit [Ping timeout: 248 seconds]
eschnett has joined #ste||ar
bikineev has quit [Remote host closed the connection]
zbyerly_ has quit [Remote host closed the connection]
zbyerly_ has joined #ste||ar
ct__ has joined #ste||ar
ct__ is now known as ct-clmsn
<ct-clmsn>
hello!
zbyerly_ has quit [Remote host closed the connection]
zbyerly_ has joined #ste||ar
ct-clmsn has quit [Quit: Leaving]
<zao>
g'morning
hkaiser has quit [Quit: bye]
zbyerly_ has quit [Remote host closed the connection]
zbyerly_ has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
taeguk has joined #ste||ar
zbyerly_ has quit [Remote host closed the connection]
zbyerly_ has joined #ste||ar
EverYoung has joined #ste||ar
zbyerly_ has quit [Ping timeout: 246 seconds]
EverYoung has quit [Ping timeout: 246 seconds]
Matombo has joined #ste||ar
zbyerly_ has joined #ste||ar
zbyerly_ has quit [Ping timeout: 276 seconds]
quaz0r has quit [Ping timeout: 240 seconds]
heller has quit [Quit: No Ping reply in 180 seconds.]
heller has joined #ste||ar
Matombo has quit [Remote host closed the connection]
quaz0r has joined #ste||ar
Vir has quit [Ping timeout: 240 seconds]
Matombo has joined #ste||ar
pree has joined #ste||ar
vamatya has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
david_pfander has joined #ste||ar
Matombo has quit [Remote host closed the connection]
<jbjnr>
I think the emauil of acceptance should have gone to you too
<bikineev>
it seems not, but anyway. Are you going to present the paper alone?
<jbjnr>
if you want to come- you're invited!
<bikineev>
heh, I think I'll be preparing a talk for meetingcpp - my first talk in English
<jbjnr>
I don't know much about this conference. I think it's a fairly small one - but a nice location I've never been to
<bikineev>
so nervous
<bikineev>
i've never been either
<jbjnr>
very strange to have co-authors that you've never met
<bikineev>
heh
<jbjnr>
the internet has changed everything
<bikineev>
in a good way
<heller>
jbjnr: yay!
<jbjnr>
this is the first paper I've written that is about HPX.
<jbjnr>
pity it's a shit conference. would have been nicer to do an IPDPS one
<heller>
Agreed, still nice
<heller>
When can I call you dr John now?
<jbjnr>
we will not know anyhtig abot that until I submit and they agree to examine it.
ajaivgeorge_ has quit [Ping timeout: 268 seconds]
ajaivgeorge_ has joined #ste||ar
K-ballo has joined #ste||ar
ajaivgeorge_ has quit [Ping timeout: 240 seconds]
ajaivgeorge_ has joined #ste||ar
eschnett has quit [Quit: eschnett]
hkaiser has joined #ste||ar
Matombo has quit [Ping timeout: 260 seconds]
<hkaiser>
jbjnr: nice job! paper accepted!
<jbjnr>
yes. thanks
<hkaiser>
jbjnr: could you send a copy of it to aserio for inclusion on the webpages?
<jbjnr>
will do. Have to do a cmaera ready version by end of month. will send him one then.
<hkaiser>
tks
<jbjnr>
hkaiser: will you be visiting europe in the near future?
<hkaiser>
nothing planned, why do you ask?
<jbjnr>
I need a signed stament from my co-authors for the PgD, I will draft something and email it to you. If you were visiting heller or anything, I could do it together ...
<jbjnr>
^signed statement
<jbjnr>
^PhD
<jbjnr>
<grrr> typing
<hkaiser>
jbjnr: let's plan for using snail mail for this
<jbjnr>
yup
Matombo has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
eschnett has joined #ste||ar
<jbjnr>
hkaiser: heller anyone else - I just received notification from our HR dept that we have to start planning projects for internships next year. This year we had shoshana and made great progress on multiple thread pools. If anyone has good ideas for projects like that, feel free to let me know.
<hkaiser>
jbjnr: GPU integration is always needed, however who knows where we will be with this next year
<hkaiser>
jbjnr: also, pls see hpx-users
<jbjnr>
I was thinking about GPU stuff myself. I need to come up with a good project/use case that would fit an internship nicely.
<jbjnr>
hpx-users. will do
<hkaiser>
jbjnr: what state has shoshana left things in?
<jbjnr>
thread pools are working, executors are going, all is fine except we have a race in shutdown (or something) that causes a segfault and we need to clean up the perf counters since there are now multiple pools and asking for idle rate (&etc) needs to work correctly on default pool, all polls, or whatever pool the user chooses
<hkaiser>
k
<jbjnr>
but essentially, it's good and we'll doa a PR soon
<hkaiser>
jbjnr: do you plan to keep working on this?
<jbjnr>
yes. I'm working on it 100% at the moment as I'm using it for raffaelel's matrix stuff and putting all comms on a separate thread pool
<hkaiser>
ok, understand - very nice!
<jbjnr>
still trying to track down scheduler bugs, but it's all related
<hkaiser>
did you solve your HP problem, btw?
<jbjnr>
not yet,
<jbjnr>
but now I'm reproducing raffaele's original problem here, so it's only a matter of time
<hkaiser>
k
<jbjnr>
(paper about thread pool executors will be forthcoming)
<hkaiser>
cool
<heller>
hkaiser: Hey, you pinged last night?
<hkaiser>
did I?
<hkaiser>
hmmm
<heller>
Yeah
<heller>
Yesterday afternoon for you
<hkaiser>
forgot what this was about ;)
<heller>
Probably reviewing prs :p
<hkaiser>
no, it was something else...
bikineev has quit [Ping timeout: 240 seconds]
<heller>
Due to family being sick and me being the only healthy, writing got delayed. Hooray...
<heller>
hkaiser: see pm please
<jbjnr>
but at least you have another paper to list in your biblio section :)
<hkaiser>
heller: the appveyor stuff seems to work nicely now - if you really want to set up a test system on that windows cluster then this could serve as a template on how to easily set things up
<hkaiser>
it's using vcpkg, which btw also supports using the microsoft mpi, would be a good incentive to get that working as well
<K-ballo>
we got appveyor ci now? cool
<K-ballo>
looks surprisingly simple
<K-ballo>
what's "Appveyor.MSBuildLogger"?
<heller>
hkaiser: appveyor: great
<hkaiser>
K-ballo: they provide that for on-the-fly error reporting on the website
<heller>
looks like it is just an adaption of mpich, should work out of the box then
<heller>
will you be in toronto next week?
<heller>
I really need a copy of andrew suttons work ;)
<heller>
K-ballo: they got me hooked indeed. This and co_await bring HPX to a whole new level
<jbjnr>
is there a talk about these metaclasses I can watch yet?
<jbjnr>
(I assume that's what you're referring to
<jbjnr>
)
<K-ballo>
I have yet to look at that proposal
<heller>
jbjnr: yes, that's what I am referring to. I just read the proposal and played around with them a bit and had some discussions with one of the authors
<hkaiser>
heller: I now remember
<hkaiser>
heller: would you be able to join the gb call today to discuss what to do with the paper?
<hkaiser>
jbjnr: same^^
<heller>
hkaiser: yes, I was planning to
<jbjnr>
what time?
<hkaiser>
great
<jbjnr>
4pm here?
<hkaiser>
30 minutes from now
<jbjnr>
ok. I join
<hkaiser>
good
<heller>
It looks like ASPLOS is the most favored venue
<hkaiser>
webex is down...
<hkaiser>
heller: can you reach the site?
<heller>
let me try
<heller>
hkaiser: No, can't get on
<hkaiser>
darn
<hkaiser>
is there any other venue we could use?
<jbjnr>
I can open a scopia room if you want
<jbjnr>
(to connect to it)
<hkaiser>
jbjnr: yes, pls
<jbjnr>
give me 2-5 mins
<heller>
nope
<heller>
jbjnr: does it work on linux?
Matombo has quit [Ping timeout: 260 seconds]
<hkaiser>
wash[m]: could you create a meeting room through LBL?
<jbjnr>
scopia 6708201
<hkaiser>
link?
<jbjnr>
works on android I think
RostamLog has joined #ste||ar
<hkaiser>
this link is a marketing site, no downloads
<heller>
or call in numbers
<heller>
the app asks me for username, passwort and server address
<jbjnr>
never mind, we can wait till next week I'm sure
<hkaiser>
we can do a phone line
<heller>
where is wash when you need him?
<jbjnr>
scopia phone dial in number is +41 43 244 89 30 | 6708201#
<jbjnr>
last part is room code
<heller>
is there a german/US one as well?
<jbjnr>
try it
<jbjnr>
no, I'm in CH
<jbjnr>
and I opened the room
<heller>
ah, ok
<jbjnr>
use office phone so they pay for the call though !
<hkaiser>
ok, I'm on the ctt phone line
<hkaiser>
I sent the details
<heller>
I am not at an office at the moment ;)
<heller>
jbjnr: the number is dropping my call
<jbjnr>
type here then we'll read it out :)
<heller>
skype to the rescue...
<jbjnr>
skype then
<heller>
no, you can make calls with skype
<heller>
but it doesn't seem to freaking work
<heller>
hmm, used to work ...
<hkaiser>
guys, let's reschedule for next week...
<hkaiser>
I hung up now
hkaiser has quit [Quit: bye]
david_pfander has quit [Ping timeout: 276 seconds]
<jbjnr>
ok. going home for a swim now
<jbjnr>
Maxime just told me we should use appear.in - website where you just create a video meeting and go. easy easy
<jbjnr>
he just showed me how it works. lovely
<jbjnr>
see y'all
<heller>
appear.in/stellar-group
<heller>
let's see
patg[w] has joined #ste||ar
patg[w]_ has joined #ste||ar
<patg[w]_>
parsa parsa[w] hello
Matombo has quit [Ping timeout: 260 seconds]
patg[w] has quit [Ping timeout: 260 seconds]
patg[w]_ is now known as patg[w]
Matombo has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
<patg[w]>
heller: I'm having some difficulty building HPX on a cluster here, it seems I need to build most dependencies myself which boost and gcc do you recommend and do I need flags just for c++14
pree has quit [Ping timeout: 268 seconds]
<heller>
patg[w]: what do you have available?
<patg[w]>
Almost anything its a test bed that they keep up to speed
<heller>
which version of gcc?
<patg[w]>
4.9.2 through 7.1.0
<heller>
ok, you can use whichever
<patg[w]>
which boost should I use
<heller>
the newest, probably
<patg[w]>
I love that last word
<patg[w]>
heller: how is the thesis going?
<patg[w]>
Oops gtg
patg[w] has quit [Quit: Leaving]
hkaiser has joined #ste||ar
<heller>
hkaiser: this appear.in seems to work very well
<hkaiser>
heller: ok
<hkaiser>
cool
<hkaiser>
works on firefox/chrome only, though
<heller>
too bad only john joined ;)
<heller>
ok
<heller>
but i reserved us the room
<hkaiser>
k, tks
eschnett has quit [Quit: eschnett]
mbremer has joined #ste||ar
mcopik has quit [Ping timeout: 276 seconds]
Matombo has quit [Remote host closed the connection]
<github>
[hpx] hkaiser created fixing_2701 (+1 new commit): https://git.io/vQwi4
<github>
hpx/fixing_2701 a7929b3 Hartmut Kaiser: Making sure Vc V2 builds without additional HPX configuration flags...
<github>
[hpx] hkaiser opened pull request #2742: Making sure Vc V2 builds without additional HPX configuration flags (master...fixing_2701) https://git.io/vQwir
pree has joined #ste||ar
parsa[w] has quit [Read error: Connection reset by peer]
Vir has joined #ste||ar
parsa[w] has joined #ste||ar
Vir has quit [Ping timeout: 240 seconds]
Vir has joined #ste||ar
Vir has quit [Changing host]
Vir has joined #ste||ar
eschnett has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
Vir has quit [Ping timeout: 240 seconds]
Vir has joined #ste||ar
Vir has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
Matombo has joined #ste||ar
akheir has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
Matombo has quit [Remote host closed the connection]
patg[w] has joined #ste||ar
<github>
[hpx] hkaiser created fixing_2743 (+1 new commit): https://git.io/vQw5j
<github>
hpx/fixing_2743 e4ccb61 Hartmut Kaiser: Adding ability to statistics counters to reset base counter...
<github>
[hpx] hkaiser opened pull request #2744: Adding ability to statistics counters to reset base counter (master...fixing_2743) https://git.io/vQwdU
<hkaiser>
heller: both things are equally important
<hkaiser>
heller: btw, what should we do with the PR related to the stack overflow?
bibek_desktop has joined #ste||ar
pat[w] has joined #ste||ar
<pat[w]>
hkaiser: I finally built HPX on that cluster
<pat[w]>
I've been busy so just got back to it last night
eschnett has joined #ste||ar
<heller>
hkaiser: I'd like to test it eventually
<hkaiser>
pat[w]: great!
<hkaiser>
heller: yah sure
<hkaiser>
heller: I was more referring to licensing issues - does thatcreate any?
zbyerly_ has joined #ste||ar
zbyerly_ has quit [Ping timeout: 255 seconds]
mcopik_ has joined #ste||ar
hkaiser has quit [Quit: bye]
mcopik_ has quit [Ping timeout: 248 seconds]
<diehlpk_work>
heller, When we do not ship the library there should be no issue or?
pat[w] has quit [Quit: Leaving]
<heller>
I didn't check
<heller>
The question is about derivative work. The gpl is very strict here
<heller>
Similar to patents
hkaiser has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
bikineev has quit [Remote host closed the connection]
<jbjnr>
when I run inside the debugger I get exception "unknown counter type /threads/count/cumulative" when I run on the command line I don't. Any idea what env flag or other might trigger this behaviour
<jbjnr>
diehlpk_work: I meant me. I didn't do anthing about writing material for book possibilities, and I really believe we need one
<hkaiser>
jbjnr: no idea :/
<jbjnr>
aha. discovery
<jbjnr>
looks like the exception happens in terminal, but is quietly swallowed
<jbjnr>
in debugger I get a breakpoint and then shutdown is so fast afterwards, I thought app was terminating
<jbjnr>
and the policy is a local var - is it going to be copied by value, or ref? looks like my high/low priorty policies are being overwritten - if they are copied by ref - that would explain it
<jbjnr>
I will dig into code and see, but if you know the answer ...please tell
<hkaiser>
hmmm
<hkaiser>
should get copied, I'd say
<hkaiser>
need to look at the code, though
<hkaiser>
jbjnr: policy is a launch policy?
<jbjnr>
nopoe. that's not it. I create a new policy on each loop, so it's not being overwritten
<jbjnr>
yes, high or low type
<hkaiser>
definitely copied
<jbjnr>
something fishy is afoot. I can see my high priorty tasks being run as high priority, but so are the low ones, so somehow all tasks are going onto the hp queue by mistake. i wondered if the policy was being reset, before the datflow actually happened and the wrong policy was being used