Rage 128 is a step into the right direction and ATI will most certainly do very well with it. Here are some reasons why:
- Best MPEG2-decoder implementation available inside a 2D/3D chip world wide. With the Rage 128 DVD decoding becomes a breeze, the CPU can do other things, e.g. take care of the stuff offered by interactive DVDs.
- Best 32-bit color 3D-rendering of all 2D/3D chips, showing the state of the art. Maybe that'll make 3Dfx wake up a bit.
- Support of 32 MB of local memory, now also done by 3Dfx and their upcoming Voodoo3.
- Available as version for motherboard implementation with 64-bit memory interface, offering a very attractive solution for onboard 2D/3D in lower cost systems. This is very interesting for OEMs. The Voodoo3 will e.g. only come with 128-bit memory interface (as far as I know), requiring a huge space on motherboards and thus making it a lot less attractive.
- Support of digital RGB out for flat panels, becoming very important in 1999.
- Support of DTV, no other 2D/3D chip is able to do this.
The Rage 128 will have to overcome the current problems and it could be a very serious competitor to NVIDIA's RIVA TNT. However, right now there's no reason for NVIDIA to become nervous. Rage 128 is not really faster than TNT, considering the results from the 'crusher' demo and the Direct3D performance. The additional features of Rage 128, the 32 MB memory support and particularly the impressive performance in 32-bit color mode could be a good reason for preferring Rage 128 over TNT, once the problems have been sorted out. Then ATI should also have an answer to TNT-2 though.
All in all I'm absolutely convinced that ATI will continue its success with Rage 128, there's hardly any doubt about it. Rage 128 plus ATI's highly successful way of doing business in the retail and particularly the OEM market are an almost unbeatable team. So it may be that you will never buy an ATI Rage 128 card in the shop, but get a complete system with it instead.