They say "The P4 3.2 wins the CPU benchmark while the Athlon 64 FX-51 runs away with the memory test with 11599 points - an all-time record!". Uh excuse me, the P4EE had 14253, where did the 11599 come from? A HT system shines in memory tests because it hides memory latency, obviously its going to smoke a single context CPU (even 64-bit).
Also the rhetoric about Intel being infantile and what not really shows the biasedness here. The one thing I completely agree with is the subject heading. A P4EE works with software today at the same (or better) performance than an Athlon 64-bit will work with the software tomorrow.
There is a reason why Intel didn't go 64-bit with 100% 32-bit backwards compatibilty. The results show it clearly. I'm sorry to say it but I think AMD has dug their own grave.
He said that because the P4EE is not out yet. IMO it is baised against AMD. First of all they are running 32 bit software in an incomplete 64 bit OS. This slows down performance, second they inclined not to include software compiled for AMD-64, like the archiving programs amd included with their test rig. i dont recall them leaving out benchmarks where SSE2 was a concern. The door should go both ways, tom did not leave out video encoding benchmarks optomised for P4s back in the day, but now he does with AMD.<P ID="edit"><FONT SIZE=-1><EM>Edited by MajinChewY on 09/26/03 00:20 AM.</EM></FONT></P>
Interesting I thought his review was running 64-bit software under a 64-bit OS. If that's the case then perhaps it's closer than I thought. In general I found the review to be really confusing. You would figure they would do a much more professional job at evaluating all the colorful graphs they put up.
First of all they are running 32 bit software in an incomplete 64 bit OS. This slows down performance,
No it doesn't. And, AMD was touting it's 32-bit compatiblity, which should not affect how it runs 32-bit apps. Incomplete 64-bit O/S? You mean, WinXP 64? According to their test setup, all systems were running WinXP, SP1...no WinXP 64.
second they inclined not to include software compiled for AMD-64, like the archiving programs amd included with their test rig.
This statement makes no sense. Why should a reviewer use a program designed by the company to test the same company's product? If A64 does what it says, it should run all available software, and not just the ones included with their test rigs.
i dont recall them leaving out benchmarks where SSE2 was a concern.
A64 has SSE2. It shouldn't be a factor.
The door should go both ways, tom did not leave out video encoding benchmarks optomised for P4s back in the day, but now he does with AMD.
So, its a reviewers fault that an application was optimized for one CPU over another, then? Well, let's just take all optimized software out and run...what? Most software has been optimized for one or the other, or both. If a programmer decides to optimize for one over another, that's not the reviewers fault. So, let's get rid of UT and Q3, any encoding, and anything else optimized for only one CPU and we are left with Sandra synthetic tests.
It's a review. If you don't like it fine. Email THG and ask why certain apps weren't run and why others were. I looked at the test setup, and it didn't look any different from alot of other reviewers.
How many watts does it take to get the center of CPU core?