System setup and notes
Hardware
- AOpen Aeolus GeForce FX5900XT 128MB, 8x AGP, 390/700
- NVIDIA GeForce FX 5700 Ultra reference 128MB, AGP8X, 475/906
- PowerColor Radeon 9600XT Bravo 128MB, AGP8X, 500/675
- ATI Radeon 9800 Pro 256MB, 8x AGP, 380/700
- AMD Athlon 64 3400+, 2200MHz
- EPoX 8HDA3+ S754 VIA K8T800 Motherboard (12/12 BIOS)
- Corsair XMS3500C2, 2 x 256MB, 2-2-2-6 @ DDR400
- Samcheer 420w PSU
- AMD reference cooler
- IBM 120GXP 40GB Hard Drive
- Dell P991 19" flat-faced CRT monitor
Software
- Windows XP Professional w/SP1
- ATI CATALYST 4.1
- NVIDIA ForceWare 53.03
- VIA Hyperion v4.51 chipset drivers
- DirectX 9.0b Runtime
- FRAPS 2.0.0
- 3DMark 2001SE v330
- 3DMark03 v340
- AquaMark3
- Splinter Cell (Beyond3D Demo)
- Unreal Tournament 2003 Retail (patched up to 2225 - HEXUS custom benchmark)
- X2: The Threat
- Call Of Duty - HEXUS Custom Benchmark
- RTCW: Enemy Territory - HEXUS Custom Benchmark
- Tiger Woods PGA Tour 2004 (patched to v1.1) - HEXUS Custom Benchmark with FRAPS
Notes
I've chosen to include a Radeon 9800 Pro 128MB card into the mix. Current pricing puts it at anywhere between £175 and £250. That puts it a notch or two above the FX 5900 XT's intended pricing, as we've seen models range from £120 - £160, but we're attempting to cover the £100 - £200 market via 4 cards. Relative performance should dictate a potential buyer's purchasing decision. Is it worth paying the extra, say, £30 or so above a Radeon 9600XT?, and is a Radeon 9800 Pro 256MB worth another £30 - £50 more than the FX 5900 XT?. That's the purpose of our investigation. Please note that the comparison ATI Radeon 9600XT isn't strictly reference material. PowerColor decided to bump up the memory speed to 675MHz, up from the default 600MHz.
2D was subjectively excellent at all tested resolutions and from either VGA or DVI outputs. The large-ish cooler's fan didn't create any undue distraction. We could barely hear it over the general noise present in a quiet-ish system. No problems to report on the installation and stability fronts. Benchmarks were carried out three times. The highest and lowest results were discarded. On to the numbers.