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 joined #ste||ar
anushi has joined #ste||ar
galabc has quit [Quit: Leaving]
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
K-ballo1 has joined #ste||ar
K-ballo1 has quit [Client Quit]
K-ballo has quit [Read error: Connection reset by peer]
heller has quit [Ping timeout: 245 seconds]
heller has joined #ste||ar
nanashi55 has quit [Ping timeout: 260 seconds]
nanashi55 has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Client Quit]
anushi has quit [Ping timeout: 260 seconds]
anushi has joined #ste||ar
Anushi1998 has quit [Quit: Bye]
Anushi1998 has joined #ste||ar
anushi has quit [Read error: Connection reset by peer]
Anushi1998 has quit [Ping timeout: 240 seconds]
Anushi1998 has joined #ste||ar
Chewbakka has joined #ste||ar
<heller>
I have to say, I really like spack
Chewbakka has quit [Remote host closed the connection]
jaafar has quit [Ping timeout: 256 seconds]
hkaiser has joined #ste||ar
Chewbakka has joined #ste||ar
churli has joined #ste||ar
Chewbakka has quit [Remote host closed the connection]
churli has quit [Quit: Leaving]
jakub_golinowski has joined #ste||ar
<jakub_golinowski>
M-ms, yt?
<M-ms>
jakub_golinowski: yep
<jakub_golinowski>
so just to make sure: if I have 8 PUs assigned to HPX and the parallel_for is called with default parallel execution then no matter the range it will be always divided into 8*4=32 chunks?
<M-ms>
so that's why I said that if you let opencv do the chunking you should force a chunk size of 1
nikunj97 has joined #ste||ar
hkaiser has quit [Quit: bye]
<jakub_golinowski>
M-ms, so what is the num_tasks parameter in there?
<M-ms>
jakub_golinowski: it should be the number of elements you're looping over
<M-ms>
each element is potentially a task
<jakub_golinowski>
Ok, thanks
jakub_golinowski has quit [Ping timeout: 256 seconds]
<heller>
M-ms: did we settle on wether we want to move to sphinx?
<M-ms>
heller: no, I've been putting it off indefinitely because I'm so unhappy with the API reference output, but I'm tempted to just ignore it and say it's good enough (because the manual is more important than the reference)
<heller>
right
mcopik has joined #ste||ar
<heller>
I am also leaning towards thinking doxygen itself might not be the right thing after all
<M-ms>
hmm, why is that? what would be the alternative?
<M-ms>
or is the alternative nothing at all/manually created docs for the API?
<heller>
yeah, manually created reference docs
<heller>
but no idea if that's feasible
<M-ms>
I guess it'
<M-ms>
s feasible, but annoying to set up the first time
<M-ms>
I guess there's a danger of docs not matching the code and so on... but to be honest it's so much more flexible
<heller>
well, that danger is already there with doxygen
<K-ballo>
do you actually mean that? how do you think it associates?
<nikunj>
K-ballo: the current implementation is only for gcc and clang
<K-ballo>
but clang only in the case where one does not have static linking?
<nikunj>
K-ballo: I didn't get what you want to say
<K-ballo>
what's the precedence of && and ||?
<nikunj>
&& first and then ||
<nikunj>
oh now I get what you're saying. Yes I should change it accordingly. I didn't didn't check it through :/
ITassin has joined #ste||ar
diehlpk has quit [Ping timeout: 264 seconds]
<zao>
nikunj: I asked a friend about running code early on Windows. There's a feature of the executable format called "TLS callbacks", which run extremely early in process startup, before the entrypoint.
ITassin has quit [Client Quit]
<zao>
While it may not be applicable for HPX-as-a-runtime, it might be good to know about.
<nikunj>
zao: Thanks, I'll look into TLS callbacks and check if I can do anything to make it work on windows.
<nikunj>
K-ballo: I have changed accordingly, please review it.