[MEncoder-users] Performance on multi core machines - a specific enquiry re avisynth mt input

[Michael Rozdoba]

RC wrote:
> On Sat, 29 Sep 2007 19:34:28 +0100
> Michael Rozdoba <mroz at ukgateway.net> wrote:
> 
>> Maybe I need to ask the MEncoder devs, 
> 
> Reimar is one of the MEncoder devs...

LOL, thanks for telling me :)

>> How is MPlayer involved?
> 
> Mencoder and MPlayer are 90%+ identical code, and the project for both
> is called MPlayer... 

Thanks for the explanation. I did wonder, as they obviously share a 
large part of their functionality.

>> Could you give me a sample MEncoder command line which takes an mp4 as
>> input & performs a computation intensive (ie cpu heavy) transcode to 
>> x264 say & which ought to run multithreaded on a quad core?
> 
> mencoder -nosound -ovc x264enc -x264encopts
> pass=(1/2):turbo=2:threads=auto:bitrate=1000 (Input) -o (Output.avi)

I did think I was being lazy & should look it up but got called away for 
tea. Thanks. I'll try this now before posting...

Hmm, seems the Megui build doesn't have x264enc... [ -ovc help ] ah, 
it's just x264...

The first pass gets the cpu to 67% (for a first pass I'd not expect 
100%, maybe around 80% tops, as it is getting through 120fps & might be 
disk bound - especially as in this case I used a huffy input).

The second pass crashes with no error (other than the windows 'send 
error report' pop up), just after outputting:
Pos:   0.0s      5f ( 0%)  0.00fps Trem:   0min   0mb  A-V:0.000 [0:0]

Ignoring the crash for now, this tells me MEncoder will use multiple 
cores. Still not sure it tells me where the problem lies though.

I don't know anything about Windows internals, but perhaps MEncoder is 
still only allowing the decoding of its input to run in a single thread. 
Is this possible?

-- 
Michael Rozdoba