News Focus
News Focus
icon url

chipdesigner

05/10/06 8:15 PM

#4855 RE: chipguy #4854

Which in practice means on average it provides much less
than 20% performance gain and in worse case probably
near zero gain.


Yeah, reducing latency to main memory by 2/3rds is nothing. And IPF is going to take over the world. And SGI is a great investment.
icon url

mas

05/11/06 5:56 AM

#4866 RE: chipguy #4854

It does not work like that. It either pays off bigtime, and there are many benches where K8 is > 20-30% better than K7, or not at all due to the work being compute/cache/bandwidth rather than latency limited. It just so happens that the former type of gain predominates otherwise AMD would be bankrupt now if it was relying on K7 type performance ;-).