New system performs worse than old system

[gschoen]

I recently built a new computer and cannot figure out why my scores for memory and 2d/3d graphics are lower than my old system. I've been using Performance Test 7.0, but got the same comparative results with 6.1. I've also tried using 3DMark06 and got lower performance but that benchmark is harder to compare. The processor scores are about 50% better, as expected so that doesn't seem to be a problem. The hard disk is the same and the score is the same, as expected. Memory mark score is about 50% lower, even using exactly the same memory chips. The systems are:

Phenom 9600 BE, stock cooler
Foxconn A79A-S Motherboard (790FX/SB750)
Corsair Dominator 2x1GB CM2X1024-8500C5D
Sapphire Radeon HD 3870
Sigma Shark 585W PSU 12v1 20A 12v2 18A
WD VelociRaptor 300GB
Windows 7 Beta 64 bit
********Old system*********
Athlon X2 6000, stock cooler
Biostar N4F AM2 motherboard
Sapphire Radeon X1950 Pro
Corsair Dominator 2X1GB Ram (same sticks)
Sigma PSU (same)
VelociRaptor (same)
Windows 7 Beta 64 bit

I swapped the cards between the computers using the same OS and drivers and the 3870 gets similar performance (within 5%) to the X1950 in the old system but I expected a performance increase. This leads me to believe there is an issue with the card or drivers seperate from a memory issue. Not to make this post too long with everything I've unsuccessfully tried but it includes clean installs, different catalyst driver versions with complete uninstalls, BIOS update, default and adjusted BIOS settings including memory timings.

My main concern is trying to determine if there is a system or hardware issue or a benchmark issue (if the scores aren't directly comparable for instance) and if it is a the system trying to determine where the problem is - drivers, BIOS settings, hardware malfunction.

I appreciate any suggestions you may have.

Gregg

[David (PassMark)]

On average the HD 3870 only does marginally better than the X1950 Pro. So I would not be expecting dramatic differences. The faster CPU should have helped the 2D and 3D test a bit however.

While it might be interesting to see the actual memory test results, I am not sure what else to suggest to fix it. You have already looked at the obvious things.

[gschoen]

I didn't see that chart, thank you that answers my question. I've concluded it is a function of the benchmark, different cards seem to do better at some benchmarks than others. For the memory scores the Phenom seems to be deficient compared to the Athlon X4, but a memory score really doesn't matter if other systems aren't affected.