|
[an error occurred while processing this directive]
|
SE: Why will some Glaze3D chips have a 128-bit external memory interface and others will only have a 64-bit interface?
SL: The architecture will support 64-bit and 128-bit external memory arrays and actually does not require any external memory. This was done to address specific segments, feature sets, and price points in the market. In most applications and resolutions, the external memory will mainly be used for texture storage, which does not require more than 1-1.5 GB/sec of bandwidth, a 64-bit bus could provide this type of speed.
SE: What speed RAMDAC will it use? What refresh rates will the Glaze3D be capable of, both in 2D and 3D?
SL: We will announce all final specifications in March.
SE: Is the Glaze3D feature set and design frozen or still a work in progress?
SL: It is completely frozen. Again we will announce all final specifications in March.
SE: Your site says that 3DGlaze uses a tried and tested rendering Architecture. Is this all software testing in your own simulations or have you licensed/bought a pre-existing rendering architecture? Is there working silicon yet? If not, when do you expect to have it?
SL: The simulators are up and running and are serving as the basis for all of our test vector and software development. We expect to be demonstrating XBA enabled silicon in 2Q.
SE: Bitboys has expressed expectations of 200fps with all features on @ 800x600x32 in Quake III Arena. What processor would that occur with? How did you arrive at the 200fps figure?
SL: Currently neither Intel or AMD have a CPU that is capable of running Quake 3 at 200 fps. But with Gigahertz processors within the next 12 months they are rapidly closing in on this number. We obtained these results from our simulator. The simulator is a cycle-accurate simulator, meaning that it works exactly like the final silicon. In fact, we use the simulator to generate all our test vectors for netlists and volume production.
SE: What is the difference between the 3DGlaze's environment bump mapping and Matrox's environment mapped bump mapping?
SL: Matrox has a good implementation of environmental bump mapping. They are closely following the Microsoft guidelines, which Bitboys licensed to Microsoft. Without going into detail on our own implementation we will have equal quality, but at a fairly significant performance increase.
|
|
|