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/
<jbjnr>
hkaiser: hkaiser just fyi - the hello world test on 2 localitiies still gives me strage results - I see more messages than threads allocated, some are doing it twice.
<jbjnr>
(sorry for strange typing, my irc client has gone odd again and I cannot see what I type)
<hkaiser>
jbjnr: ok - I was hoping that heller's fixes would fix this as well
<jbjnr>
will look tomotrrow. goodnight
<hkaiser>
g'night
teonnik has quit [*.net *.split]
teonnik has joined #ste||ar
<wash>
hkaiser: ping
<wash>
K-ballo: ping as well
<hkaiser>
here
<wash>
We don't implement the future blocking in the destructor (when created by async) - right?
<hkaiser>
yes
* K-ballo
lurks
<wash>
Err, sorry - yes, we don't implement it?
<hkaiser>
we don't implement it
<wash>
gotcha
<hkaiser>
we can't have runaway threads
<wash>
My understanding from Olivier is that the reason it blocks is because we do not want to have runaway std::threads
<hkaiser>
nod
<wash>
What's so bad about runaway std::threads?
<hkaiser>
we wait for all threads to run to completion before exiting
<wash>
yah
<wash>
But let's say we wanted std::future to not block too. What's bad about a runaway thread?
<hkaiser>
ask hans
<hkaiser>
he forced the blocking into std::future
<wash>
Welp
<wash>
Hans is not at the next meeting >.>
<wash>
(I'll send him a mail)
<hkaiser>
wash: nice
<hkaiser>
would you mind giving me access to the paper itself?
<wash>
yah
<wash>
github invite sent
<hkaiser>
tks
<wash>
hkaiser: the version in the repo is outdated
<wash>
the slides are a little less outdated, but still outdate
<wash>
d
<hkaiser>
k
<wash>
I'll send an update as soon as we can. The design we have now I think you'll be happier with. It is API-backwards compatible
<hkaiser>
nice
<wash>
And I believe it is semantically compatible with what HPX does, but not the standard.
<wash>
gotta run, back later
wash has quit [Quit: leaving]
teonnik has quit [Ping timeout: 260 seconds]
wash has joined #ste||ar
wash has quit [Quit: Lost terminal]
<K-ballo>
hkaiser: did you see that article about forceinline?