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/
eschnett has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
anushi_ has quit [Quit: Bye]
hkaiser has quit [Quit: bye]
eschnett has quit [Quit: eschnett]
nanashi55 has quit [Ping timeout: 244 seconds]
nanashi55 has joined #ste||ar
jaafar has quit [Ping timeout: 250 seconds]
parsa[w] has quit [Read error: Connection reset by peer]
<simbergm>
heller: I definitely want to do a release soon
<simbergm>
did you have something in progress that you want to see in the release
<simbergm>
?
<simbergm>
jbjnr: did you see the hpx backend got merged into opencv?
<jbjnr>
yes. I saw that
<jbjnr>
very good
<jbjnr>
simbergm: why are you so desperate to release?
<jbjnr>
OCD?
<simbergm>
jbjnr: of course
<simbergm>
not desperate, just want it to become a regular thing
<jbjnr>
ah. ok
<jbjnr>
I can see the merit in that.
<simbergm>
or maybe I am without knowing it...
<jbjnr>
a regular release cycle is a good idea, then we don't try to rush/panic features into a release, because we know another one will come along soon.
<simbergm>
yeah, that's more or less the intention
<simbergm>
most users can wait 6 months for a new release with whatever feature they happen to need
<heller>
simbergm: I'd like to get lazy init and the future allocation into the next release
heller has quit [Read error: Connection reset by peer]
<simbergm>
does it make sense to have 3146 without stack reuse?
<heller>
no
hkaiser has joined #ste||ar
K-ballo has joined #ste||ar
hkaiser has quit [Quit: bye]
<K-ballo>
jbjnr: your logic is flawed, boost libraries have made it into the standard while still holding obvious boost dependencies
aserio has joined #ste||ar
<zao>
Silly silly libraries :)
hkaiser has joined #ste||ar
eschnett has joined #ste||ar
<jbjnr>
K-ballo: explain please? what std libraries depend on boost?
<K-ballo>
that's exactly the flaw in the logic
<K-ballo>
boost libraries with boost dependencies were adopted into the standard, the standard does not become dependent on boost because of it
<K-ballo>
once something is adopted into the standard it becomes a responsibility of implementors to provide all the bits, dependencies included
<K-ballo>
the existing libraries show that the spec is implementable, it doesn't really matter how or which piece builds on which
<K-ballo>
standard proposals today still depend on boost half the time
<K-ballo>
maybe what remains unclear is that the standard is a specification, not an implementation
<zao>
I look forward to the <web_*> hell.
<hkaiser>
zao: I like your offcie door
<zao>
^^
<jbjnr>
K-ballo: yes. I understand what you meant now. Quite right, but it's not me you need to convince unfortunately. Once all those boost libs are in std, then we're done, but until then, we're stuck with them
<hkaiser>
jbjnr: there never will be a time that all boost libs will be in boost
<zao>
The standard needs to be implementable on anything that compilers target. A lot of Boost is extremely narrow in scope, particularly anything that depends on third party libraries or particularities of the OS or platform.
<zao>
If you pull in everything under the sun, you end up with the behemoths that are web browsers.
<K-ballo>
we don't need to wait, not for all of them, impmementations are never part of specifications
<K-ballo>
the standard doesn't care if we use boost::thread or std::thread internally, it can't even tell
<heller>
We should just submodule the boost libs we need and include them in our build system. Then it looks like we're free of boost...
aserio1 has joined #ste||ar
david_pfander has quit [Ping timeout: 246 seconds]
aserio has quit [Ping timeout: 250 seconds]
aserio1 has quit [Ping timeout: 240 seconds]
aserio has joined #ste||ar
aserio has quit [Ping timeout: 240 seconds]
jaafar has joined #ste||ar
<heller>
hkaiser: i'd be available from now on
<hkaiser>
heller: let's skip today, sorry
<hkaiser>
heller: let's reconvense next week :/
<hkaiser>
reconvene
<heller>
hkaiser: ok
<heller>
hkaiser: then let's talk tomorrow
<hkaiser>
ok, sounds good
hkaiser has quit [Quit: bye]
kisaacs has joined #ste||ar
<kisaacs>
Hey everyone, I'm generating traces of various sizes as vis test data -- usually I run LRA and ALS (instrumented versions with breast_cancer and MovieLens) with -n 1, -t2 which isn't a whole lot of parallelism. Do you have recommended options for something more in the tens-of-threads range to start off? Thanks.
<K-ballo>
guys, stop fighting on github threads please