Re: SSE performance library
Re: SSE performance library
- Subject: Re: SSE performance library
- From: Ries Brad <email@hidden>
- Date: Tue, 5 Sep 2006 18:21:29 -0600
On Sep 5, 2006, at 5:04 PM, Shawn Erickson wrote:
On 9/5/06, Ries Brad <email@hidden> wrote:
Thanks. The docs are a little dated. Everything related to this in
the Reference Library is dated 2005-09-08 and refers exclusively to
PPC. Ran a simple test though, and clearly it's tuned for SSE.
<http://developer.apple.com/documentation/Performance/Conceptual/
Accelerate_sse_migration/index.html>
and
<http://developer.apple.com/documentation/MacOSX/Conceptual/
universal_binary/universal_binary_vector/chapter_6_section_1.html>
O.K., I suppose one conclusion from the docs above is that the
Accelerate framework is optimized for SSE. However these docs don't
actually say that. Only that developers should use the Accelerate
framework for vector-oriented processing. I don't think it's too
much to ask that the actual document for the library I'm interested
in make mention of support for SSE.
http://developer.apple.com/documentation/Performance/Conceptual/vDSP/
vDSP_Library.pdf
Given how it talks exculsively about Altivec and that it is supported
on the G3, a reasonable conclusion is that optimized for Intel means
that the default scalar implementation is used.
Regardless, I see what I need is optimized for SSE.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden