<The only reason I slipped that 550 benchmark in there was to tone down some of the hype that the K7 has drawn. There are many non-believers, and it's about time for the market to start believing that AMD does have a very competitive chip, even in performance.>
Well, this e-mail confirms that "seventh-generation technology" is only "very competitive" with sixth-generation technology.
Of course, Anand really isn't telling us the whole story, and that is confusing a lot of people out there. We don't know what was "defeatured" in that K7. We don't know how fast the off-chip L2 cache was running, and this is a major factor in Winstone 99 scores even with that oversized 128K L1 cache. We don't know how fast the processor bus was running (100, 133, or 200 MHz?), not that it makes much of a difference anyway. And most importantly, we haven't seen any tests which exercise the K7's floating point or 3D performance.
I'd sure like to know what the heck is keeping the K7 from being the great Pentium III killer that the hype made people believe. Remember when AMD was saying that the K6-III was positioned against the Pentium III, and the K7 was going to be positioned against Willamette? Seems like the AMD hopeful will have to settle with "very competitive" instead of "mind-blowing performance."
But hey, there's always the MHz, right? As usual, I'll believe it when I see it.
Tenchusatsu |
