hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar-group.org | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | This channel is logged: irclog.cct.lsu.edu
diehlpk_work has joined #ste||ar
diehlpk has joined #ste||ar
K-ballo has quit [Quit: K-ballo]
diehlpk has quit [Quit: Leaving.]
hkaiser has quit [Quit: Bye!]
diehlpk_work has quit [Remote host closed the connection]
diehlpk has joined #ste||ar
diehlpk has quit [Client Quit]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
nanmiao has quit [Ping timeout: 256 seconds]
Yorlik has joined #ste||ar
<rori[m]> nice :D
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
<hkaiser> gnikunj[m]: yt?
Yorlik has quit [Ping timeout: 260 seconds]
<gnikunj[m]> hkaiser srinivasyadav227 rori hey, sorry for missing out on today's meeting :/
<hkaiser> gnikunj[m]: should I buy you an alarm clock?
<gnikunj[m]> I think that's a good present for my bday
<hkaiser> +1
<gnikunj[m]> But I slept at 7. So, I had to give up waking up at 9 too xD
<hkaiser> makes perfect sense
<gnikunj[m]> This compiler project has me and my sleep covered these days ;_;
diehlpk has joined #ste||ar
diehlpk1 has joined #ste||ar
diehlpk has quit [Ping timeout: 264 seconds]
diehlpk1 has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
Yorlik has joined #ste||ar
diehlpk has quit [Client Quit]
diehlpk has joined #ste||ar
diehlpk has quit [Client Quit]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
nanmiao has joined #ste||ar
diehlpk has joined #ste||ar
diehlpk has quit [Client Quit]
diehlpk has joined #ste||ar
diehlpk has quit [Quit: Leaving.]
K-ballo has quit [Quit: K-ballo]
K-ballo has joined #ste||ar
Yorlik has quit [Ping timeout: 245 seconds]
diehlpk has joined #ste||ar
diehlpk has quit [Client Quit]
diehlpk_work has joined #ste||ar
FunMiles has joined #ste||ar
diehlpk_work_ has joined #ste||ar
diehlpk_work has quit [Quit: Leaving]
diehlpk_work_ has quit [Remote host closed the connection]
diehlpk_work_ has joined #ste||ar
<FunMiles> I am on MacOS and am trying to initialize HPX from inside my app. Until a recent update, it was working the way I wanted, but now it is forcing -e,_initialize_main through CMake. Is there an easy way to avoid this flag?
<FunMiles> The other thing I am having trouble with is that the thread back-off system does not quiet down. As soon as the app start, even if it is idle on a terminal input read, I see between 50% and 90% of one core CPU running.
<hkaiser> FunMiles: that's how it ever has been, but you can force the scheduler to step back faster
<hkaiser> for the linker flag, I need to refer to ms[m], he might have an answer on how to avoid it
<hkaiser> (I'm on windows where this is not an issue)
<hkaiser> hold on, nimight be able to help as well
<FunMiles> hkaiser: The docs give some flags. But since I am not controlling the start of HPX like I could when the -e flag is not activated, I don't know how to set the backoff values.
<hkaiser> FunMiles: do you explicitly call hpx::init()?
<hkaiser> or do you #include <hpx/hpx_main.hpp>?
<FunMiles> No. My main.cpp, as it is not intended to deal with HPX directly, is not including hpx/hpx_main.hpp. I followed an example that talks about dynamically loaded libraries.
<FunMiles> It sets up a thred_local variable that keeps registry with HPX.
<hkaiser> FunMiles: how do you initialize hpx, then?
<FunMiles> That thread_local object constructor's calls hpx::start(start_function, __argc, __argv, init_args) with made-up arguments.
<hkaiser> ahh
<FunMiles> Then hpx::register_thread(rts_, name); is called and the converse is done in the destructor.
<hkaiser> the init_args should allow you to pass any command line arguments you need
<FunMiles> I can find where that example was in the source code of HPX.
<FunMiles> examples/quickstart/init_globally.cpp
<hkaiser> nod
<hkaiser> FunMiles: that example sets some command line arguments as well
<FunMiles> Yes, but since with the -e, flag the initialization of HPX has been taken over, do the arguments passed then affect the run? or is that hpx::start igoned? It does return true, so it does not seem to create a conflict.
<FunMiles> I have not tried to pass actual arguments that would have an effect yet.
<hkaiser> not sure
<FunMiles> Is there a way to change those parameters after the start?
<hkaiser> no
<hkaiser> start initializes hpx, those have to be known at that point
<FunMiles> Then I'm screwed until I can remove the -e, flag :D
<hkaiser> pls ask ms[m], he'll be around tomorrow (European time)
<FunMiles> *self sarcastic smile*
<FunMiles> I will, thanks.
<FunMiles> It is possible that eventually, all my use of HPX will be in a dynamically loaded library, so that flag issue will be gone, but for now, I'd rather continue with the current setup.
K-ballo has quit [Remote host closed the connection]
K-ballo has joined #ste||ar
<hkaiser> gonidelis[m]: yt?
<gonidelis[m]> hey
<hkaiser> see pm, pls