[Webkit-unassigned] [Bug 74842] SSE optimization for FFTFrameFFMPEG.cpp::multiply()

Tue Dec 20 18:14:37 PST 2011

https://bugs.webkit.org/show_bug.cgi?id=74842

--- Comment #5 from xingnan.wang at intel.com  2011-12-20 18:14:37 PST ---
(In reply to comment #4)
> great to hear about the Biquad optimization!
> 
Now I am doing some work on the Biquad optimization, I will submit another patch when finishing it.
> I'll have to look more closely, but I'm pretty sure we only need the zvmul() version which operates on non-interleaved (planar) complex data like we use in FFTFrameFFMPEG.  I *thought* that was the same as the one used in FFTFrameMac.

Yes you are right , sorry for misunderstanding the vDSP_zvmul, which surely uses non-interleaved complex data. I`ll update the patch soon.

BTW, do you think it makes sense to add two functions like vDSP_ctoz and vDSP_ztoc in VectorMath? Maybe we can improve the performance of packing and unpacking complex data after some optimization on them.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.