NASA Benchmarks Power Mac G5

[Hammer]

It seems that NASA didn’t use the latest i875 chipsets (i.e. dual channel PC3200 DDR SDRAM and 800Mhz QDR FSB)....

[Hammer]

With a G5 optimisation, the G5/2Ghz should match the P4/3.2 Ghz or even beat it by 20%

One problem, one has to factor in the i875 800Mhz QDR FSB (current product release) and Prescott Core (future product release).    

There is no bench for P4 which does not have a vector unit

Pentium 4 does have SIMD unit(s), except it's broken in some cases. It’s works with most entertainment titles and media encoding/decoding applications.
This was highlighted by AMD around the launch of K8 Opteron (around April 2003). AMD also uses Portland Group’s X86-32 complier to show this case.

Focusing on Mflops/mhz ratio is nice but it’s useless** in the product release vs product release comparisons. **It’s purely academic.

I don’t know why NSAS has picked Pentium 4** when the real competitor to the PPC G5 should be  AMD’s Opteron/Athlon 64.  **Until Intel releases a “cheaper than Itanium” desktop/workstation 64bit CPU solution, the Pentium 4 is not a competitor in the 64bit market.

Note that Portland Group’s X86-32 complier will undergo product revision e.g. Version 5 (currently under beta testing). Refer to http://www.theinquirer.net/?article=10292

[Hammer]

next month's release of Athlon-64 on same technology mobos as G5 should prove a realistic benchmark comparisons.

Note that “Athlon 64” is not a multi-processor capable.  AMD64 K8 product line doesn’t map 1-to-1 with PPC 970 product line.

Refer to
http://www.theinquirer.net/?article=10353
“Athlon 64” may move to Opteron’s 940 pin socket (initially). Socket 754 may go like Socket 6 (Pentium Classic) or Socket 423(Pentium 4) i.e. the market dictates socket adoptions.