[MPlayer-dev-eng] [PATCH] use timer-lx.c on darwin & osx

[Nicolas Plourde]

On 26-Apr-05, at 1:55 PM, Arpi wrote:
>
> i'm against.
> maybe as an option...
> timer-osx uses a much more precise timer, than linux's usleep
> thing which dosnt work far as accurate on osx as on linux.
> (on linux it's mostly using /dev/rtc to do the job, but on
> osx it doesnt exists so it will fallback to inaccurate usleep)
>

I ran some benchmark using both timer and my result are showing
not much performance lost using timer-darwin.c(not taking cpu load  
into account).
But has previously reported by Dieter Shirley <dete at mac.com>,  
usec_sleep is called way too often.

quote:
>> The problem, somewhat ironically, is that the Darwin usleep()  
>> timer is
>> really quite accurate, much more accurate than the 10ms that mplayer
>> seems to assume.  Here's the code from mplayer.c:2440:
>>
>>          float min=softsleep?0.021:0.005;
>>    current_module="sleep_timer";
>>    while(time_frame>min){
>>    if(time_frame<=0.020)
>>    usec_sleep(0); // sleeps 1 clock tick (10ms)!
>>    else
>>    usec_sleep(1000000*(time_frame-0.020));
>>    time_frame-=GetRelativeTime();
>>
>> What ends up happening with this code is that usec_sleep(0) gets  
>> called
>> *hundreds* of times for each frame of video.  This isn't a trivial  
>> call
>> either, since it touches the kernel.  The upshot is that sleeping can
>> take up to 30% of the CPU, more than decoding the video!

here is the "shark" report (original code):

- 22.4% ml_set_interrupts_enabled (mach_kernel)
- 8.3% __memcpy (commpage [libSystem.B.dylib])
   7.7% shandler (mach_kernel)
+ 7.6% nanosleep (libSystem.B.dylib)
| - 7.6% usec_sleep (mplayer_darwin_20)
- 4.7% altivec_yuv2_bgra32 (mplayer_darwin_20)
- 3.5% synth_1to1 (mplayer_darwin_20)
- 2.3% msmpeg4_decode_block (mplayer_darwin_20)
- 1.7% __gettimeofday (commpage [libSystem.B.dylib])
- 1.5% int2float (mplayer_darwin_20)
- 1.4% gldDestroyQuery (ATIRadeon8500GLDriver)
- 1.4% idct_add_altivec (mplayer_darwin_20)
- 1.4% put_pixels16_altivec (mplayer_darwin_20)
- 1.1% __fixunsdfdi (libSystem.B.dylib)

and now patched with Dieter recommendation:

- 15.5% __memcpy (commpage [libSystem.B.dylib])
- 8.2% altivec_yuv2_bgra32 (mplayer_darwin_05)
- 5.8% synth_1to1 (mplayer_darwin_05)
- 4.0% msmpeg4_decode_block (mplayer_darwin_05)
- 3.5% ml_set_interrupts_enabled (mach_kernel)
- 2.8% put_pixels16_altivec (mplayer_darwin_05)
- 2.5% gldDestroyQuery (ATIRadeon8500GLDriver)
- 2.3% idct_add_altivec (mplayer_darwin_05)
...
+ 0.1% nanosleep (libSystem.B.dylib)
| - 0.1% usec_sleep (mplayer_darwin_05)

suggestion?

====
Nicolas Plourde <nicolas.plourde at sympatico.ca>