In our review of the Radeon X1600 XT reference card manufactured by ATI, we came to the conclusion that “it is hard to describe Radeon X1600 XT since from our results it looks like a completely schizophrenic graphics chip”, and one thing that could explain the odd behavior of Radeon X1600 XT would be the driver we used, that was still a beta version.

With Radeon X1600 XT from HIS we used the latest official stable driver from ATI (Catalyst 6.1) and still the results were absolutely the same: the model from HIS achieved exactly the same results (the be honest, ATI reference card was slightly faster one some tests, but the advantage of this card was below 4%, so we must consider that there is no significant performance difference between them).

It is faster than mid-range chips like GeForce 6600 GT, Radeon X600 XT and Radeon X700 Pro, but when compared to GeForce 6800 GT, GeForce 6800 GS, Radeon X800 GT and Radeon X800 GTO it is a mess.

On 3DMark03, which is based on Shader 2.0 (DirectX 9.0), GeForce 6800 GT was between 22.82% and 57.01% faster, GeForce 6800 GS was between 23.50% and 50.88% faster and Radeon X800 GTO was up to 20.79% faster than Radeon X1600 XT. Radeon X1600 XT achieved a performance similar to HIS Radeon X1600 XT under this software, with a slight advantage of up to 4.14% on some tests.

On 3DMark05, which is based on Shader 3.0 (DirectX 9.0c), Radeon X1600 XT was between 7.30% and 15.85% faster than GeForce 6800 GT and achieved a performance similar to GeForce 6800 GS on almost all tests, but on 1600x1200 with image quality settings enable Radeon X1600 XT was 15.75% faster. Radeon X1600 XT was between 12.69% and 36.67% faster than Radeon X800 GT and between 8.62% and 16.28% faster than Radeon X800 GTO.

On Far Cry, Radeon X1600 XT was faster at 1024x768 and 1280x1024 with no image quality settings enabled than GeForce 6800 GT and GeForce 6800 GS. On these resolutions it achieved a performance similar to Radeon X800 GT and Radeon X800 GTO. On other video configurations, however, GeForce 6800 GT was between 7.82% and 51.96% faster, GeForce 6800 GS was between 11.76% and 66.24% faster, Radeon X800 GT was between 8.40% and 16.51% faster and Radeon X800 GTO was between 19.80% and 41.80% faster.

On Doom 3, excluding 1600x1200 with and without image quality settings enabled, where it achieved a performance similar to Radeon X800 GT, Radeon X1600 XT was beaten by GeForce 6800 GT, GeForce 6800 GS, Radeon X800 GT and Radeon X800 GTO in all our tests.

So, excluding 3DMark05, the “old” Radeon X800 GTO is faster than the new Radeon X1600 XT! You may think that’s because 3DMark05 is Shader 3.0. Can be, but how to explain the better performance of Radeon X800 GTO on Far Cry, a Shader 3.0 game?

O.k., let’s assume that Radeon X1600 XT is really the best Shader 3.0 mid-range graphics chip, but how about the other kind of games? What is the point of having a terrific Shader 3.0 performance while its performance on all other kind of games is lower than its competitors?

If you compare pricing, it is even more insane: models based on Radeon X1600 XT with 256 MB can be found between USD 167 and USD 200, while a Radeon X800 GTO with 256 MB is being sold today between USD 185 and USD 210.

So unless you run exclusively DirectX 9.0c (Shader 3.0) games (which we think is quite improbable), Radeon X800 GTO is a better choice than Radeon X1600 XT.