sandpile.org -- discussion forum

discussion forum

message

Name:	rbkrf
eMail:	rbkrf@go2.pl
Date:	April 23, 2004 at 07:17:42
Subject:	Re: differences in EM64T vs. AA-64
In Reply To:	Re: differences in EM64T vs. AA-64 by The Owl on April 22, 2004 at 07:03:18
Text:	\| http://marc.theaimsgroup.com/?l=linux-kernel&m=107766481408468&w=2 \| Well, I managed to get my hands on the Microprocessor Report, and based on this report I can say, Jun (in the link you provided) isn't fully right.Acording to the Microprocessor Report the key differences are:CMPXCHNG16B is supported by the EM64T (32bits version - CMPX-CHG8B supported by both, probably CMPXCHNG16B will be supported by AMD in future CPU steppings) SYSCALL & SYSRET in 32 bits are supported by the AMD64 (in 64bits supported by both) SYSENTER & SYSEXIT in 64bits are supported by the EM64T (in 32bits supported by both) LAHF & SAHF in 64bits is supported by the AMD64 (Intel claims that those instructions generate invalid opcode on latest AMD64 CPUs, anyway - invalid on EM64T - due to some manual disorder made by AMD) and the last one - 32-bit mode physical memory = 1TB by AMD64, 64GB by EM64T As you can see the Microprocessor Report omits some of the info stated in the provided link (FPU state, other things), and vice-versa (LAHF&SAHF case), AMD64 probably will support CMPXCHNG16B, so we must wait until "the final CPU stepping" of both architectures to judge the differences. I hope that Christian Ludloff will point the differences in EM64T vs. AMD64 implementations of AA-64 on his website (maybe he is waiting for the EM64T CPUs to show up in stores?). And thanks go2 The Owl for the additional link. If anyone knows something that wasn't said above (or/and in this topic) please share your tech-info!

post a followup message
(Be nice... or be blocked. Be technical... or be erased.)

`Name:`		`optional link title:`
`eMail:`		`optional link URL:`
`Subject:`		`optional image URL:`
	Insert line breaks by hand when only about one inch remains at the right side. Otherwise your message will be unreadable.
`Text:`	\|\| http://marc.theaimsgroup.com/?l=linux-kernel&m=107766481408468&w=2 \|\| \| Well, I managed to get my hands on the Microprocessor Report, and based on this report I can say, Jun (in the link you provided) isn't fully right.Acording to the Microprocessor Report the key differences are: \| \| CMPXCHNG16B is supported by the EM64T (32bits version - CMPX-CHG8B supported by both, probably CMPXCHNG16B will be supported by AMD in future CPU steppings) \| \| SYSCALL & SYSRET in 32 bits are supported by the AMD64 (in 64bits supported by both) \| \| SYSENTER & SYSEXIT in 64bits are supported by the EM64T (in 32bits supported by both) \| \| LAHF & SAHF in 64bits is supported by the AMD64 (Intel claims that those instructions generate invalid opcode on latest AMD64 CPUs, anyway - invalid on EM64T - due to some manual disorder made by AMD) \| \| and the last one - 32-bit mode physical memory = 1TB by AMD64, 64GB by EM64T \| \| As you can see the Microprocessor Report omits some of the info stated in the provided link (FPU state, other things), and vice-versa (LAHF&SAHF case), AMD64 probably will support CMPXCHNG16B, so we must wait until "the final CPU stepping" of both architectures to judge the differences. I hope that Christian Ludloff will point the differences in EM64T vs. AMD64 implementations of AA-64 on his website (maybe he is waiting for the EM64T CPUs to show up in stores?). And thanks go2 The Owl for the additional link. If anyone knows something that wasn't said above (or/and in this topic) please share your tech-info! \|


	Note: The above eMail form fields may look unaligned if you are using a browser other than Netscape Navigator version 3.0.

currently posted followup messages
(You may have to press the RELOAD button of your browser.)