> >Although video encoding is one of the places those extra general > >purpose registers in AMD64 mode can show themselves, > > that must be in bizarreo world. Every time I have run encoding tests > with AMD CPUs against Intel they perform slower by factors of three, > and four. >
Not AMD vs Intel, x86 vs x86-64 (AMD64). AMD did, after all, author the 64-bit extensions to the x86 instruction set. Some choose to refer to the lot as AMD64--including Microsoft, btw. And yes, the extra GPRs do seem to have a positive impact on encoding performance. > > >The part that'll sting a bit for you is that 3.06Ghz Netburst Xeons > >will get kicked around by a 1.86Ghz Core 2 Duo, the Xeon box isn't a > >clunker, but... well, Netburst always was a bit rubbish :p > > I have a P4p Xeon and I am not sure it is, or isn't a Netburst ,as it > was purchased after Intel dropped Netburst. However, I don't buy your > analysis anyway. Every encoding test I have done, over the last four > years, blows away anything AMD has put out. I have never seen any > legitimate source claim that AMD could stand up to Intel when it > comes to video editing and encoding. Anything Pentium 4 or Pentium D (or Intel Xeon prior to 3xxx, 51xx, or 72xx models) uses the Netburst marchitecture.
