Test System, AquaMark, SPECviewperf, 3D Mark 03
This content was originally featured on Amdmb.com and has been converted to PC Perspective’s website. Some color changes and flaws may appear.
|Processor||XP 1800+ @ 166×10 (1660 MHz)|
|RAM||512MB PC2700 (2-2-2-6T-Fast)|
|Operating System||Windows XP + SP1|
|NVIDIA Driver||Detonator v52.14|
|ATI Driver||Catalyst 3.8|
Aquamark 3 has become one of the popular alternatives to 3DMark. This benchmark uses the Krass Engine which is used to power Massive Development’s Aquanox 2 game. The benchmark uses many DirectX 9 features and therefore is a good measurement of DX9 capabilities.
Looking at the above chart, the FX5700 Ultra performs very well. It significantly beats the older FX5600 Ultra and matches the Radeon 9700 in DirectX 9 performance. As we have heard from NVIDIA in the past, the series 50 Detonator drivers are designed for DirectX 9 specifically and increases FX-series performance dramatically. So far the FX5700 Ultra looks like to be a worthy competitor, unlike the FX5600 Ultra which never came close to being in the same class.
SpecPerf uses industry standard applications (like 3DS Max and Pro/Engineer) to create a benchmark that measures real-world professional performance. This benchmark is purely OpenGL based and gives a good sense of how graphics hardware performs in rendering applications. For gamers, this benchmark would give you an idea of raw performance, but is a poor reflection of actual game play.
Here we see some interesting results. In some tests, the FX5700 Ultra actually matches the FX5900 Ultra. This is very good news for you professionals out there who do a lot of CAD/CAM. For half the price, you can get performance seen in the more expensive FX5900 Ultra. Comparing the FX5700 Ultra to the soft-modded Radeon 9500 (our “Radeon 9700” surrogate) we see that the NVIDIA card is better at some applications and worse in others. The FX5700 Ultra appears to be a good alternative to the Radeon 9700 depending on the application.
3DMark 03 has been at the center of some controversy in recent months. Despite this, it remains a good method of comparing results between systems assuming you are comparing apples to apples. 3DMark 03 uses a combination of DirectX 7, DirectX 8, and a small sampling of DirectX 9.
|GT1 – Wings of Fury (FPS)||102.1||126.0||100.7
|GT2 – Battle of Proxycon (FPS)||16.4||28.7||17.6
|GT3 – Troll’s Lair (FPS)||15.7||
|GT4 – Mother Nature (FPS)||18.4||24.8||11.6
|Fill Rate (Single-Texturing, MTexels/s)||8541.9
|Fill Rate (Multi-Texturing, MTexels/s)||10897.0||1985.1||1065.4
|Vertex Shader (FPS)||13.0
|Pixel Shader 2.0||24.1||37.4||15.0
The FX5700 Ultra has a significant 600 point lead over the FX5600 Ultra thanks to higher core/mem clocks and DDR-II. What I find most significant here are the drastically improved shader performance numbers. Poor shader performance has been the center of some discussion with previous generation FX products, however it appears that NVIDIA’s next generation shader (CineFx 2.0) gives their cards the much needed boost.
Comparing the “Radeon 9700” to the FX5700 Ultra, we see that the Radeon 9700 beats the FX5700 Ultra in 7 of 9 tests.