[Webkit-unassigned] [Bug 75528] Optimize the multiply-add in Biquad.cpp::process

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=75528


Raymond Toy <rtoy at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rtoy at google.com




--- Comment #64 from Raymond Toy <rtoy at google.com>  2013-07-09 10:43:54 PST ---
(In reply to comment #63)
> I might be completely wrong (I just happened to skim past this patch), but shouldn't the SSE2 code be a runtime check?  The #ifdef SSE2 is only good for knowing if you can use the intrinsics or not (if the compiler supports SSE2).  I don't know the current standing expectation of having SSE2 support is, but last I remember it wasn't considered a requirement:
> 
> https://groups.google.com/a/chromium.org/forum/#!topic/chromium-dev/66p8sJtluJA

I think you are correct.  It's been a while since I looked, but I think on Linux and Windows the compiler flag that enables SSE2 support is not set, so the SSE2 code is never compiled in.  On Mac, SSE2 has always been available on x86 machines and I think SSE2 code is compiled by default.  

There should be no issues with this code.  In any case, this particular patch was rolled out.

Ideally, we would like a runtime check and use the SSE2 code when available, falling back to the C reference when SSE2 is not available.

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