Graphics card history: How pixels learned to run

Graphics card history: How pixels learned to run

A new beginning?

nVidia countered again with a new graphics card. The GeForce3 entered the market and hit the stores with the high price of DM 1000 (approx. € 510). The performance of the GeForce3 was not that great at the beginning, since DirectX-8 games were not available and the drivers did not support the new features apart from the shaders. Only in high resolutions and in 32-bit color depth could you clearly set yourself apart from the GeForce2ultra. Nevertheless, vertex and pixel shaders promised gigantic performance boosts, which were mainly documented by the extremely TnL-heavy 3DMark 2001. nVidia itself spoke of up to 7x higher performance compared to the GeForce2, which can of course only be reproduced in scenarios that appear very constructed. The promise of revolutionized graphics was owed to us to this day, as appropriately adapted games are still almost not available at all.

nVidia Geforce3 from Asus

From its shadowy existence, ATI has now stepped back into the limelightand with the Radeon8500 presented a worthy opponent of the GeForce3 after the? GeForce2 killer? could not do justice to the high praise he had received in advance due to the higher frequency. The tiresome driver problem was also an issue here.

ATi RadeonSDR
ATi Radeon8500

Not only was the Radeon faster than the GeForce3 in many areas, but ATI was also able to perform well in terms of image quality and the features it offered convince. Just like the GeForce3, the Radeon8500 had a programmable T&L unit (pixel and vertex shader). Ahead of the GeForce3, the Radeon8500 had two interesting features. SmoothVision, a higher quality, but slower anti-aliasing variant (which unfortunately was not available at the beginning and still does not seem to be error-free today) and Truform, which should round off low-poly models in 3D games without much effort, but a patch the programmer needs it.


It is not yet possible to predict where the future will take us in terms of graphics cards, but it is certain that the course towards photorealism will continue . Even if we have only just finished the announcements and first tests of the next generation of graphics cards, the next step will be extremely interesting, as we seem to be approaching a new threshold in the realism area.

GeForce4 TI cards from Asus

The two leading 3D standardsDirect3D from Microsoft and OpenGL are currently in the final stages of the current versions and the already announced successors promise to bring even more innovations with them than any version before them.

But this time it seems as if the upcoming standards will be interpreted a little more far-sightedly than the triple steps that have prevented us from enjoying the new graphics so far, since hardly any game developer has them could or wanted to take part in whole intermediate steps. Even the pixel and vertex shaders introduced with DirectX 8 are apparently (see DirectX 8.1) just a first step towards the fully programmable GPU, which then really earned its name for the first time.

Let's hope the market will Calm down a little and this insane pace of 'little innovations' is replaced by healthy, far-reaching progress.

That would benefit both the industry and the customer.

For questions is as always our forum is responsible. Furthermore, we collect high-quality photos of old graphics cards in an extra thread . Everyone is welcome to present their old gems there.

This article was interesting, helpful or both? The editorial team is happy about any support from ComputerBase Pro and deactivated ad blockers. More on the topic of Ads on ComputerBase .