[Webkit-unassigned] [Bug 77950] SSE optimization for vsvesq and vmaxmgv

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Feb 7 15:22:08 PST 2012


https://bugs.webkit.org/show_bug.cgi?id=77950





--- Comment #2 from Chris Rogers <crogers at google.com>  2012-02-07 15:22:07 PST ---
(From update of attachment 125790)
View in context: https://bugs.webkit.org/attachment.cgi?id=125790&action=review

> Source/WebCore/platform/audio/VectorMath.cpp:439
> + 

nit: extra blank line - please remove here and elsewhere in similar places in this file

> Source/WebCore/platform/audio/VectorMath.cpp:440
> +        // If the sourceP address is not 16-byte aligned, the first several frames (at most three) should be processed seperately. 

spelling nit: "seperately" -> "separately"

please fix here and elsewhere in the file

> Source/WebCore/platform/audio/VectorMath.cpp:448
> +        // Now the sourceP address aligned and start to apply SSE. 

grammar nit: I'd re-word this comment as:

// Now that sourceP is aligned, use SSE.

> Source/WebCore/platform/audio/VectorMath.cpp:450
> +        float* endP = (float *)sourceP + n - tailFrames; 

shouldn't use ordinary cast here.

Can't we make endP "const float":

const float* endP = sourceP + n - tailFrames;

> Source/WebCore/platform/audio/VectorMath.cpp:462
> +        float* groupSumP = reinterpret_cast<float*>(&mSum); 

nit: I think you can use "const float*" instead of "float*"

> Source/WebCore/platform/audio/VectorMath.cpp:487
> +

nit: extra blank line

> Source/WebCore/platform/audio/VectorMath.cpp:488
> +        // If the sourceP address is not 16-byte aligned, the first several frames (at most three) should be processed seperately.

nit: "seperately" spelling

> Source/WebCore/platform/audio/VectorMath.cpp:495
> +        // Now the sourceP address aligned and start to apply SSE.

nit: see comment above about re-wording

> Source/WebCore/platform/audio/VectorMath.cpp:497
> +        float* endP = (float*)sourceP + n - tailFrames;

nit: see comment above

> Source/WebCore/platform/audio/VectorMath.cpp:509
> +        // Caculate abs of min, then campare with max.

typo: "Caculate" -> "Calculate"
typo: "campare" -> "compare"

Is this math correct??

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



More information about the webkit-unassigned mailing list