Testing the Effects of NVIDIA Detonator v44.03 Performance Increase

UT2K3 Image Quality: Detonator 44.03 Examined

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

(aka. “There’s no such thing as a free lunch.”)

Author’s Update:

In waiting for this article to go “public” on AMDMB.com, [H]ardOCP has published an article dealing with a similar issue with the v44.03 drivers (see here). I would just like to acknowledge the fact that their article, though similar, are completely different experiments and different ideas on the same driver.

I strongly encourage readers here to read through our article here and also read [H]ardOCP’s article.

Author’s Update #2 (July 22, 2003):

It has come to my attention that Beyond3D has a thread that pre-dates our investigation here on AMDMB. Though our own research was done independently of what was being done over there, I feel it important to give recognition to their hard work as well since it provides a foundation of what we’re discussing here.

So props goes to Dave Baumann and gentlemen over at Beyond3D for their great work. You can catch their discussions here.

On NVIDIA’s website for the latest v44.03 Detonator driver, it states the following:

NVIDIA's v44.03 driver download page excerpt.
“Delivers an incredible 30% increase in performance over previous Detonator 40 drivers.**

**Performance results based upon GeForce FX 5600 Ultra running UnrealTournament 2003 Antalus flyby.”

For many of us, it’s always great when manufacturers improve on their existing product by performing some tweaking and optimizations of their drivers. Unreal Tournament 2003 being a very popular game (and a very popular benchmark), it shows favourably on NVIDIA that they are committed to improving gaming for the community.

However, when writing another article for AMDMB (soon to be published), I noticed something peculiar. Using the v44.03 driver, the FX5600 and the Ti4800SE are nearly equally matched in practically all 1024×768 tests … except Unreal Tournament 2003 where the FX5600 beats the Ti4800SE by a very significant margin (when anisotropy is turned on). I found it curious that it is in only UT2K3 the FX5600 “trounces” the Ti4800SE, whereas in other tests, the Ti4800SE more often superior.

Could it be that NVIDIA’s v44.03 “30% increase in performance” adjustment is causing this irregularity? Is somehow this performance tweak only affecting UT2K3 and not any other game? If this tweak only applies to UT2K3, does it come at a price of something else?

It is our intention today on AMDMB, to answer these questions and test the effect of NVIDIA’s “30% increase in performance”.

« PreviousNext »