Re: dword ptr in Intel x86 assembler
Re: dword ptr in Intel x86 assembler
- Subject: Re: dword ptr in Intel x86 assembler
- From: Steven Kortze <email@hidden>
- Date: Sun, 14 Aug 2005 21:58:38 -0400
Eric,
In the x86 world, there are two different flavors of assembly
syntax. "dword ptr" is Intel's assembly syntax. The GNU tools
(gcc, GNU as, etc.) use AT&T syntax. You can find a quick primer
on the differences between Intel and AT&T syntax at <http://
www.imada.sdu.dk/~kslarsen/Courses/dm18-2005-spring/Litteratur/
IntelnATT.htm>.
Just as important as this is that gcc's inline assembly syntax is
very different from Visual C++'s syntax. In particular, gcc
requires that you tell it which registers you're changing.
I also found this link <http://www.codeproject.com/cpp/gccasm.asp>,
when I was looking for information. It describes the AT&T vs Intel
differences. Thanks for the other link.
I'd strongly suggest using the intrinsics. They'll work out best
in the long run, and you'll be able to have only one version of
your code for all x86 platforms.
The x86 asm routines are good to study for effective use of
registers. After some trial and error (make that many trials and many
errors), I did get some good performance from the SSE intrinsics.
I've been looking at Intel's literature and also the information from
the SSE page on the Velocity Engine area of the Developers website.
Thank you for the info and suggestions,
Steve
_______________________________________________
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