On 09/03/2014 12:32 AM, Tobias Doerffel wrote: > Hi, > > first profiling results seem to be a bit disappointing - even though > they probably do not say much about RT improvements. I rendered > unfa-spoken with stable-1.1, master and memmgr branches and plotted > the resulting files using gnuplot:
Hm, that's weird. I didn't expect much performance gain at this point but if there's actually loss then that's a problem... There are some places where the memmgr curve looks more even, though - could be maybe those are spikes caused by allocations in earlier branches? In any case, the branch does need some optimization. We need to figure out what's causing this performance hit... Can looking up sequential chunks in an array really be more CPU-heavy than allocating the memory dynamically? Can there be other factors in play here - there are some other changes in the same branch... mostly very small ones though. ------------------------------------------------------------------------------ Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ LMMS-devel mailing list LMMS-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lmms-devel
