What prompted this rant(Apologies beforehand)
The new Apple workstations were released today. I think Apple goofed big time in their benchmarks. They compare their system with serial ata to Dell systems with parellel ata (regular ol' ide) also, the specs were differently skewed. here are the specs as reported by the independent testing house, veritest
(When veritest didnt specify specs, i used mfgr defaults)
Apple G5
PPC970 2ghz 512KB cache x2
1.5GB DDR333
ATI radeon 9700 64mb DDR
Seagate serial ata Barracuda 120gb
Dell Dimension 8300
32bit 3.06ghz Hyperthreaded p4 512k cache 800mhz FSB
2GB DDR400
200gb IDE 7200rpm ata100?
ATI radeon 9800 pro 128mb DDR
i875p Chipset
Dell Precision 650
32bit Xeon 3.06 512k cache 533mhz fsb x2
2GB DDR266
120gb 7200 ata100
Ati Radeon VE
Both of the dell's could have been configured with faster hardware. the Apple hardware used in the benchmark is not available for purchase through Apple (120gb Seagate) they use Western Digital drives. Furthermore, faster systems exist, available from dell or homebrew/custom manufacturing. SATA has been available on the PC front for a ong while now, though its performance has not justified replacing standard IDE drives. Apple did not test a system with like components, nor what the PC can be configured at for performance. The Dimension system has Serial ATA readily available as a raid setup (Greatly improving performance) but Apple chose to go with a single IDE drive in both PC setups. In this case the drives were not even the maximum bus available, ata133 or sata. Faster CPU's exist for both the Dimension and the Precision system, 3.2ghz CPU's to be exact. Furthermore, the Precision Workstation 360 is configurable with the utmost of PC hardware available at the moment.
I find it cheap that Apple chose to benchmark the highest end Apple hardware available with the upper middle hardware from the PC section. I would love to own Apple hardware, and i hope my next workstation is a G5. However i am embarrased and disapointed that they had to make such a low blow with their benchmarks. Perhaps this is indicative of the real world performance of the g5, or other factors that i do not know of during the benchmark process. Surely an out of the box redhat install (both dell systems) versus a hand-held apple engineer tuned G5 is a like comparison. Unfortunately this is a Marketing department spoiling what should be a triumphant day for Apple. Numerous people have pointed out already that benchmarks are not sufficient for determining real world performance so i hope that independent system reviewers will pick up apple's dropped ball and bring a fair test to the table.