Yep, I agree with this assessment. That is, assuming Intel ships on time, etc. etc. etc. 6 months from now.
I don't think anyone has doubted this for a while - the Centrino pretty much showed this could happen last year, Intel just seemed so married to the P4 that they left their best horse in the stable for so long. CW -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Greg Sevart Sent: Friday, March 10, 2006 8:01 AM To: The Hardware List Subject: Re: [H] Were Intel's benchmarks (slightly) rigged? >I saw yesterday where Anandtech had an updated comparison that used an > updated BIOS for the AMD. They also corrected an error where the Intel > came > out ahead by 40% or so. After the correction, Intel is only ahead by 20% > or > so. > > Sorry, I do not have a link for you. > I posted said link on this list yesterday. "Only" 20% is very substantial. If the numbers Anand reported are completely accurate, then we should fast forward to the competitive landscape in 6 months to see where things lie. Let's say AMD releases the FX-62*, maybe FX-64 at 2.8 and 3.0GHz, respectively. Let's also say these chips are AM2/DDR2. Even if these enhancements are enough to catch up that 20% (which they won't be), that still leaves the best-of-breed AMD chip as an equal to Intel's entry level 2.6GHz Conroe. That's assuming that Intel makes no more improvements to the core, AND isn't accounting for the fact that the Conroe was tested with DDR2-667, not DDR2-800 like it will be at release. The 3.3GHz XE Conroe (with, I believe, twice the cache) will best anything AMD can throw at it. * The benchmarks were ran against a 2.6GHz FX-60 that was overclocked to 2.8GHz, and therefore should be indicative of how a DDR1 FX-62 would perform.
