NASA G5 study
This came across my desk last night at about 3am and I just now got a chance to really go through it. It's a comparison of the new 2GHz G5 processor to the older G4 and X86 Xeon's.
For the purposed of comparing the G5 to the xeon, it pretty much fell out as I thought it would considering the compiler issues and the G5's fairly nascent architecture.... especially considering the types of tests being run.
What I found most interesting was this:
Benchmarks from the scalar version of Jet3D are shown in Figure 1 (MFLOPS) and Figure 2 (MFLOPS normalized by MHz). In terms of raw MFLOPS, the 2GHz G5 is about 32% faster than the 2GHz P4, 97% faster than the 1.25GHz G4, 142% faster than the 1GHz G4, and within 1 MFLOP of the 2.66GHz P4.
I really hated Apple's 1.25GHz & 1.42GHz towers, they just weren't good machines compared to the 1GHz machines. Too many tradeoffs just for the MHz, and this sorta shows that off.
If clock speeds were equal, these machines would do a good job of stomping on the Xeon's. There's supposed to be a 3GHz G5 part shipping within a year, but by next July Intel really should be shipping a 4GHz+ part. It'll still narrow the race somewhat, so it should be interesting.
Posted by drunkenbatmanJuly 03, 2003, at 12:50 PM
