System Setup and Notes
HEXUS.note
The Radeon 1800 XT was initially benchmarked at the ATI press event for the X1000-series launch in Ibiza on the 30th September to the 2nd October. HEXUS took the full system described below to Ibiza and tested Radeon X1800 XT with it, without ATI intervention or interference. Upon returning to the UK on the 2nd of October, all of the tests conducted in Spain were rerun to verify correctness in a familiar testing enviroment, with more tools available to HEXUS to do so. HEXUS are presenting the UK results in this article.
Further, HEXUS are fully aware the majority of retail GeForce 7800 GTX examples usually have higher clocks than the NVIDIA set reference clocks. HEXUS believes that testing at the IHV's own stated reference clocks is the correct methodology to use for preview evaluations of new graphics chip configurations. Retail reviews after the initial performance evaluation at the IVH's own reference clock levels are then used to determine the relative performance and value of any retail SKUs an IHV's board partners are allowed to create.
| ATI Radeon X1800 XT Preview Test System | |
| Processor | AMD Athlon 64 FX-57 2800MHz, 1MiB L2 |
| Mainboard | ASUS A8N-SLI Deluxe |
| Memory | 2 x 512MiB Corsair XMS3200 XL DDR-400 2.0-2-2-5 @ 400MHz |
| Disk Drive | 80GB Seagate IDE |
| Graphics Cards | ATI Radeon X1800 XT 512MiB (R520) (625/750) NVIDIA GeForce 7800 GTX 256MiB (G70) (430/600) PCI Express 16X |
| Operating System | Windows XP Professional, SP2 |
| Mainboard Software | NVIDIA nForce4 Platform Driver Version 6.39 |
| Graphics Driver Software | ATI CATALYST Beta - 8.173.1-050921 NVIDIA ForceWare Release 75 - 78.03 |
Notes
The Radeon X1800 XT was benchmarked against GeForce 7800 GTX using a variety of popular and recent games and game demos, at resolutions up to 1600x1200 on 20.1" LCD display hardware. Benchmark results were recorded three times and the median taken as the reported result. If any result deviated by more than a couple of percentage points, all results were discarded and rerun.Image quality was examined in depth and the results are contained in the companion Technology Discussion. Testing went well throughout, barring incompatibilities with R5-series's ability to multisample FP rendertargets and certain software titles. Those will be mentioned on a title-by-title basis.
As always, and we don't mention this enough as standard text, if the game or application offers its own control for antialiasing or anisotropic texture filtering, that's always used. Otherwise the driver control panel is called upon to apply the required settings. Driver defaults for texture filtering optimisations are used unless otherwise stated.
Need to know anything else about our testing procedures, methodologies or demos used? Drop by the HEXUS.community in the 2D/3D section and let us know.
