<< It's pretty clear that our P4 owners have "bigger than your's" disease. P4 puts T-bird into shame ? Let's all belive that syntetically measured memory bandwith and Quake3 are the only thing that counts ? Oh yeah, that 3DMark's score is very meaningful, too... not What if we take some benchmarks that doesn't make use of SSE/SSE? P4 and integer performance: RC-5 client runs faster on Duron600 than on 1.5Ghz P4. Who should be in shame now? FPU performance: in 3D-graphics and raytracing P4 looks even much crappier than lower-MHZ P3 and is in same level with Duron700-800. In MP3-encoding 1.5Ghz P4 gets beaten by 1.1Ghz Athlon using non-SSE optimised Lame encoder. Using SSE/3DNow optimized encoder T-bird@1.2Ghz is still faster. Overall performance: P4 is considerably slower in linux kernel compilation and under MS C++. Where's all that memory bandwith now ? Compability: when coming out, P4 even didn't run linux with standard kernel (was it 2.2.16 back then?) and needed patch. >>
Well the biggest arguement NOS will definately mention will be future SSE2 optimization that would completely RULE!!!!!!! <-- note the NOS effect. However, I really believe that if the P4 was so superior, even without the SSE it would perform very well, but it doesn't seem the case. Before P4's debut I was expecting much greater benchmark results due to its clock speed and architectural improvement. The P4 seems to have took a step back, without SSE2 optimization it performs slower than a P3. Intel didn't solve the FPU issue(with the Athlon stomping it all over), which they once boasted back in the P2 age.