RE:"we already have more CPU power than was intended by God."
Craig,
National appears to have recognized the fact that increased integer CPU power is of little value for the vast majority of current and future users. The most common uses for home computers are:
1. Information (internet, email, etc.)
2. Games
3. Home finance
Category 1 performance is limited by communications speeds beginning at the data servers, through the local network, over the internet, over the phone line, and through the modem. Everybody would like faster communications, but increasing CPU power will do nothing to resolve this issue. As Fuchi correctly points out - a 133 MHz Media GX does as well as a PII 450.
Category 2 performance is limited by FPU processing speed and the speed of the graphics processor. NSM is behind here, but MXi was architected to address both of these issues and should be a competitive product.
Category 3 performance requires no more than a MediaGX 133.
The big performance issues of the future are going to be:
1. Battery life (CPU's run very slowly without electricity.)
2. Real time graphics processing and visualization.
3. Real time Java processing.
Brian Halla talks frequently about 9-12 hour battery life on notebooks. PC on a chip is the obvious solution for this problem. Intel will not be able to compete against the fully integrated system solution National is conjuring up.
High end graphics is not being addressed by any of the major CPU manufacturers, as far as I can tell. Perhaps the Intel/SGI or Intel/HP alliances will give Intel the edge here.
High speed, real time Java compilation/execution performance will require multiple CPU's. Intel is well positioned with Merced. Cyrix appears to have no interest. Sun Microsystems undoubtably has plans here. Pravin's suggestion that Cyrix pursue Java performance is well taken.
Scumbria |
