I haven't looked at that bench session in detail, but since even slower MPs scored much better on several other sites I think there was a configuration error with the Tyan board.
Eh, and great idea to use PS4 from the old Ziff-Davis suite as bench. That's probably working on a 2 MByte picture and isn't even MMX optmised...
November 18, 2001 4:44:03 AM
Most of what they tested was meant for single processor systems, thus the single processor one should be a bit faster. One thing i don't understand is how they got results for a single processor mp. I know they said they emulated something, but i don't see how the emulations should be compared.
Also, tyan does offer a motherboard with scsi, raid, and etc. it's 500 bux (still cheaper than the 900 dollar xeon one). Overall, i'm still curious about the dual athlon mp and am still thinking of what to purchase when i make another system athlon xp or athlon xp mp.
My biggest question is whether there i will feel a difference in performance running many applications at once (ie. internet xplorer, half life, direct connect, and etc.). Will the performance be increased with one more processor, or will it stay the same because the programs weren't optimized for dual cpu support. I would think the chipset or windows (since i am running win xp pro and it does support 1-2 processors) would split the data needing to be processed. You think anyone can help me out with my last question?
One of the major problems with that review is that it's mostly SPEC benchmarks. P4's have always done extremely well in SPEC benchmarks, yet the results usually just aren't reflected in real-world benchmarks. It doesn't help much that many of the tests really weren't SMP-oriented, and the Xeons are only 1.7GHz.