News Focus
News Focus
icon url

KeithDust2000

03/28/04 11:31 PM

#30006 RE: CombJelly #30005

Combjelly, you used 90mm^2 as the die size for the 512kb L2 cache part @ 90nm (Winchester). However, AMD already gave 102mm^2 for that one at the latest presentation !?!



icon url

Grimes

03/28/04 11:42 PM

#30007 RE: CombJelly #30005

cj how would you rate the performance of this (hypothetical) chip aaainst a top of the line pentium M?

in 'overall' peformance?
in power consumption and heat output?




icon url

8-/

04/06/04 5:18 AM

#30620 RE: CombJelly #30005

CJ, Beats me. I realize the rating for the HPaq whatever-it-is seems high, but I can't explain it. I doubt if they removed 64 bit capability, that would be a huge task for dubious benefit. To realize any IPC increase would mean pretty much a complete re-design which just doesn't make sense for that product.

How about Hypertransport added to Athlon.

8-/
icon url

j3pflynn

04/06/04 6:27 AM

#30624 RE: CombJelly #30005

CJ, Petz - I'm relatively clueless here, but if they were to do this(castrate K8), wouldn't it be more likely that they would just force one of the legacy modes or disable the long mode? At least initially, with die changes made in the future if there were substantial size/cost savings to be had?
Paul
PS. This still seems like a bizarre product to me. To castrate it on capabilities rather than performance seems unwise.