This content was originally featured on Amdmb.com and has been converted to PC Perspective's website. Some color changes and flaws may appear.

The holidays is a scarcely
a few days away and I know many of you will are looking closely at some graphics
hardware. Since most of you can't afford the premium prices of a Radeon 9800XT
or a GeForce FX5950 Ultra (see our Asus
Radeon 9800XT/TVD review for a comparison of the two), you'll probably
be looking at a mainstream card. But for $200 USD, which card do you get?
Today we will take a look at the performance of these two chipsets so you
can make sure the diamond you think you're buying isn't a lump of coal.

System Setup

• Asus
  Radeon 9600XT/TVD: The Radeon 9600XT we are using today is manufactured
  by Asus. I won't go into details about the features of this card since
  you can read about all about it in the Asus
  Radeon 9800XT review - the contents and features of both cards are
  identical.
• NVIDIA
  FX5700 Ultra: The FX5700 Ultra we are using is the reference design
  from NVIDIA. This card is very plain and there are no features to evaluate
  besides the performance aspect of it.

Synthetic
Tests

Synthetic
Tests is what we've all come to know as a "benchmark". They are
fabricated graphical events that are used to simulate a consumer's
experience with graphics hardware. However, synthetic benchmarks are a poor
representation of what is being experienced in reality, but are useful for
making good concrete comparisons of cards on similarly configured systems.
Tests like Aquamark 3, and Unreal Tournament 2003, though
based on real games and real game engines, are synthetic since the events
are fabricated and never controlled by the user.

	Aquamark 3 (Massive Development)		3DMark 03 v340 3DMark 2001SE (Futuremark)
	ViewPerf 7.1 (Standard Performance Evaluation Corporation)		Unreal Tournament 2003 (Digital Extremes, Atari)

 

Game
Performance Tests

The results
in this section of the review are taken from actual game play. FRAPS was
used to measure the frame rate as I played each game on each piece of hardware.
We've been doing this for a long time here and I'm happy to see more hardware
sites spending the time to do the same.

However,
because the numbers are recorded in actual in-game sessions, the numbers
recorded may not be what you've come to expect from a standard benchmark
(i.e. synthetic test). There will be instances where the numbers recorded
don't follow a "nice" trend line or exhibit some minor deviations.
This is to be expected since each test run on a card in any given game
will not be 100% identical to previous tests (like in a synthetic benchmark).
We endeavor to reproduce the best results in any given instance and minimize
these anomalies.

	Age of Mythology: The Titans (Ensemble Studios, Microsoft)		Freelancer (Digital Anvil, Microsoft)
	Halo: Combat Evolved (Bungie/GearBox, Microsoft)		Homeworld 2 (Relic Entertainment, Vivendi Universal)
	Jedi Knight: Jedi Academy (Raven, LucasArts)		The Elder Scrolls III: Morrowind (Bethesda Softworks, Ubi Soft)

 

Subjective
Tests

Subjective
tests is where I use my own eyes to compare image quality for different
settings on a given piece of hardware. They are subjective because the
comments and conclusions I make are based on my own personal experience
and may not reflect what you perceive. This is why we provide links to
the images used so you can judge for yourself.

Anti-aliasing quality: Morrowind

Texturing quality: Halo