2009/5/27 strk <s...@keybit.net> > On Wed, May 27, 2009 at 11:39:26PM +0800, Wei Cao wrote: > > hi, all, I tested u2_has.swf on my windows laptop, hope it's not so late > . I > > just tried the first 400 frames with the command > > > > ./sdl-gnash.exe --max-advances 400 u2_has.swf > > > > The gnash-0.8.5 with the patch last time I sent uses 61.770s CPU time, > > 51.006s is consumed in draw_shape_character , there are total 84384 > calls > > to AGG. > > > > The optimized version uses 44.744s CPU time, 36.268s in > > draw_shape_character, there are only 32123 calls to AGG now, which means, > > 5/8 calls to AGG and 25% CPU time is saved by cache. > > Visible effects on FPS (max/min/avg) ?
yeah, this is comparision of five to ten seconds gnash 0.8.5 Effective frame rate: 3.94 fps (min 3.94, avg 4.25, max 4.89, 20 frames in 4.7 secs total) Effective frame rate: 4.06 fps (min 3.94, avg 4.21, max 4.89, 25 frames in 5.9 secs total) Effective frame rate: 3.91 fps (min 3.91, avg 4.17, max 4.89, 29 frames in 7.0 secs total) Effective frame rate: 3.97 fps (min 3.91, avg 4.14, max 4.89, 33 frames in 8.0 secs total) Effective frame rate: 4.00 fps (min 3.91, avg 4.13, max 4.89, 37 frames in 9.0 secs total) Effective frame rate: 4.03 fps (min 3.91, avg 4.12, max 4.89, 42 frames in 10.2 secs total) cache version Effective frame rate: 5.98 fps (min 5.65, avg 5.81, max 5.98, 30 frames in 5.2 secs total) Effective frame rate: 5.48 fps (min 5.48, avg 5.75, max 5.98, 36 frames in 6.3 secs total) Effective frame rate: 5.73 fps (min 5.48, avg 5.75, max 5.98, 42 frames in 7.3 secs total) Effective frame rate: 5.78 fps (min 5.48, avg 5.75, max 5.98, 48 frames in 8.3 secs total) Effective frame rate: 5.75 fps (min 5.48, avg 5.75, max 5.98, 54 frames in 9.4 secs total) Effective frame rate: 5.76 fps (min 5.48, avg 5.75, max 5.98, 60 frames in 10.4 secs total) > > > > One question, I found there are three lt-sdl-gnash processes executing at > > the same time when I play a single flash file on a arm like system, one > > takes about 70% CPU and 25% the second, and 2% the third, when I kill one > of > > the three processes, the other quit too. I use dynamic linking. > > > > The question is, why there are three processes when I execut the libtool > > generated script sdl-gnash only once, where and when they are forked in > the > > code? > > It could be your process table lister considers each thread as a > separate process. One thread is the SDL sound consumer, one is the parser > and one is the main one. The parser is likely the 2% one, the sound the 25% > and the rest is main (VM + renderer). thx, I will shutdown sound thread and have a try tomorrow. > > > NOTE: current trunk fails building statically, I hope it's going to > be fixed soon for more detailed profiling. > > PS: Any progress about cleanup/port to trunk ? Need help ? Hm.. Recently I wrote a alpha blending function to blend cached RGBA bitmap to agg's screen use SIMD instructions. The cache code is stable now I think, memory is freeed when character::destroy is called, every thing is OK except after running a long time ( loop for hundreds of times ), then process's memory footprint will begin to expand from a reasonable value, e.g. 20M, to hundreds of megabytes all of a sudden, I havn't found the reason. > > > --strk; > > Free GIS & Flash consultant/developer () ASCII Ribbon Campaign > http://foo.keybit.net/~strk/services.html /\ Keep it simple! >
_______________________________________________ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
