Shortly
before the Computer 98 in Cologne, VillageTronic announced an add-on board for their PicassoIV graphics board,
consisting of a Voodoo 3D hardware accelerator. To justify development, they need a certain
amount of preorders.3DFX's Voodoo chip set has set new standards in 3D accelerators in both Performance and features. For a long period, it was the reference platform for 3D hardware until surpassed by it's own successor.
For an application to use these features, an application programmer's interface is required. This is the Warp3D project's main goal. It does this by supplying a hardware independent API. To run applications supporting the Warp3D API on a new hardware, a new driver must be written for that hardware.
We would like to emphasize that a Warp3D driver will be available as soon as the Voodoo board is released. That way, the ridiculous situation of a 3D hardware being available without driver support would be prevented. Every Warp3D application will support the hardware, making a lot of programs readily available. This includes OpenGL applications, which will automatically work on every 3D hardware.
The development of a Warp3D Voodoo driver is very easy, because there's already an existing hardware abstraction layer (Glide). The only thing to do is to map Warp3D functions to Glide functions, a job that can be achieved very easily. The Voodoo hardware has nearly all the features needed by OpenGL, which would be very interesting to people wanting to use OpenGL in future work.
This combination of top hardware and high power driver system would bring the Amiga a big step forward in the games market, there could be games running at speeds unimaginable on current Amigas, not even with he fastest processors available. The slow bus system is still the bottleneck, and with the Voodoo hardware, this can be bypassed.
All this could soon be reality, the only thing that could still prevent it would be a lack of interest, resulting in too few preorders. We want to encourage people thinking about getting such a board to preorder it *now*. Time is of the essence! More orders are required and the time of decision is drawing near.
We think this project to be very important for the future of games development on the Amiga. If the project should fail, the chances for special 3d hardware optimized games will be very small, because the Permedia2 will be available for really fast games, and the CyberVisionPPC can not be used with turbo boards from other manufacturers. The ViRGE used on the CyberVision3D can be used for low resolution games, but will not be able to handle e.g. OpenGL-based games.
Due to the slow bus transfer speed, it will not be possible to have fast games in high resolutions running at acceptable frame rates. For a resolution of 640*480*15, considered low end on other systems, a frame rate of about 10 frames, probably less on complex scenery, is the expected maximum on an Amiga without 3d hardware. Even a fast CPU is not that important, the bottleneck is the graphics boards bus. A CPU that's twice as fast will not accelerate a game by factor two, probably only by 1.5, or even less in high resolutions. On an Amiga, 3d hardware will have more impact than on other systems.
Without 3d hardware, overall graphics quality is much lower, because software rendering can't use all of the effects available to 3d hardware, including: