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
hkaiser has quit [Quit: Bye!]
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 244 seconds]
K-ballo1 is now known as K-ballo
K-ballo has quit [Quit: K-ballo]
K-ballo has joined #ste||ar
Yorlik has joined #ste||ar
<ms[m]> > <@pansysk75:matrix.org> Why is it important that the chunk size is a power of 2?
<ms[m]> pansysk75: more often than not, that helps align chunks to cache line boundaries which helps with performance, but the effect will likely vary from algorithm to algorithm
<ms[m]> >
<ms[m]> that 2-4 chunks per core is however very conservative and likely needs to be updated, especially for the executors/schedulers that aggregate chunks and only spawn one hpx thread per core
hkaiser has joined #ste||ar
<pansysk75[m]> Aha ok, makes sense, thanks a lot!
hkaiser has quit [Quit: Bye!]
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 265 seconds]
K-ballo1 is now known as K-ballo
hkaiser has joined #ste||ar
<pansysk75[m]> Also, I need a logging facility to monitor stuff internal to HPX
<pansysk75[m]> I got the HPX logging output working (with --hpx:debug-hpx-log), but I can't find an example of how to actually log stuff
<pansysk75[m]> hkaiser Would appreciate pointing me to an example when you find the time
<hkaiser> pansysk75[m]: sure
<hkaiser> pansysk75[m]: use LHPX_(lvl, " [some unique string]") << "your message"; where lvl is one of https://github.com/STEllAR-GROUP/hpx/blob/master/libs/core/logging/include/hpx/logging/level.hpp#L46-L56
<pansysk75[m]> should be plenty useful, thank you!
hkaiser has quit [Quit: Bye!]
hkaiser has joined #ste||ar