sandpile.org -- discussion forum

discussion forum

message

Name:	Belthor
Date:	March 29, 2002 at 03:30:08
Subject:	Re: Yamhill
In Reply To:	Re: Yamhill by Christian Ludloff on March 28, 2002 at 22:10:56
Text:	\|\| The patent refers to a linear address extension mechanism of a 32-bit \|\| processor. Their response to AMD's x86-64 should be more sophisticated \|\| than this. The cpu should have 64-bit integer registers and directly use \|\| 64-bit addresses in order to compete with AMD's x86-64 cpus. \| \| IA-64. :)By the way, IA-64 is not Intel's response to AMD's x86-64. Yamhill is supposed to be that. Do IA-64 cpus compete with AMD's x86-64 cpus in terms of x86 code performance? Where did you get this from? The discussion was about an alternate solution to IA-64, code-named Yamhill, that would compete with AMD's x86-64 in terms of both legacy code performance and 64-bit extensions, so that the cpu's microarchitecture will serve the legacy 32-bit code and the new 64-bit code equally well. IA-64 is not an extension to IA-32, it is a totally different architecture and the IA-64 cpus have to interpret the IA-32 instructions which results in pathetic legacy performance. IA-64 and IA-32 require different cpu microarchitectures in order to get the best code performance. Two microarchitectures in the same chip is too expensive and in order to keep the compatibility, they sacrificed the performance of the legacy code. Personally, I find IA-64 cpus big stupid monsters with huge resources who have to be told what to do, IA-32 and x86-64 cpus retarded children who desperately try to rearrange the pieces of the puzzle and Transmeta cpus some small entities with greater intelligence than their size who strive to deeply analyze the nature of the code and get the most out of their limited resources. `Belthor`

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:`	\|\|\| The patent refers to a linear address extension mechanism of a 32-bit \|\|\| processor. Their response to AMD's x86-64 should be more sophisticated \|\|\| than this. The cpu should have 64-bit integer registers and directly use \|\|\| 64-bit addresses in order to compete with AMD's x86-64 cpus. \|\| \|\| IA-64. :) \| \| By the way, IA-64 is not Intel's response to AMD's x86-64. \| Yamhill is supposed to be that. \| \| Do IA-64 cpus compete with AMD's x86-64 cpus in terms of x86 code \| performance? Where did you get this from? The discussion was about an \| alternate solution to IA-64, code-named Yamhill, that would compete with \| AMD's x86-64 in terms of both legacy code performance and 64-bit \| extensions, so that the cpu's microarchitecture will serve the legacy \| 32-bit code and the new 64-bit code equally well. \| \| IA-64 is not an extension to IA-32, it is a totally different architecture \| and the IA-64 cpus have to interpret the IA-32 instructions which results \| in pathetic legacy performance. IA-64 and IA-32 require different cpu \| microarchitectures in order to get the best code performance. \| Two microarchitectures in the same chip is too expensive and in order to \| keep the compatibility, they sacrificed the performance of the legacy code. \| \| Personally, I find IA-64 cpus big stupid monsters with huge resources who \| have to be told what to do, IA-32 and x86-64 cpus retarded children who \| desperately try to rearrange the pieces of the puzzle and Transmeta cpus \| some small entities with greater intelligence than their size who strive \| to deeply analyze the nature of the code and get the most out of their \| limited resources. \| \| Belthor \|


	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.)

Re: Yamhill by Christian Ludloff, 13:45:08 2002/03/29 (0)