On Tue, Apr 14, 2009 at 7:34 PM, Kyle Sluder <kyle.slu...@gmail.com> wrote: > On Tue, Apr 14, 2009 at 7:26 PM, Michael Ash <michael....@gmail.com> wrote: >> That has to be the most confusing technical document I've ever seen. >> I'm not certain, but after reading it about three times, I'm pretty >> sure that it's discussing data alignment on IA-64 (Itanium), not >> x86-64. Microsoft certainly thinks that x86-64 supports unaligned data >> access, with speed penalties: > > EM64T == x86_64, so I'm not sure *what* Intel is trying to say.
Yeah, I know, but they refer to both, which is what's so confusing. As far as I can tell they're talking about writing code on Itanium, and then using the fact that your Itanium code is now 64-bit clean as an easy jump over to EM64T. I could be wrong. >> And a quick test confirms that my Mac Pro, at least, has no trouble >> reading a misaligned int pointer. > > Could be GCC's codegen giving you a helping hand. I don't have an > assembler handy on this machine (except GNU as, but I don't know AT&T > syntax) so I can't try it. >From looking at the disassembly of my compiled test program, gcc is only generating a single load instruction. I think. I'm not exactly an x86-64 assembly guru. Mike _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com