CompuServe Thread

Unless there is some way of easily running the Amiga software that exists today on a RISC based system as you propose, the machine would have a hard time surviving. Without the software, they'd have to start from scratch and it would be very difficult. Steve

With the 68040 and upcoming 68060 CPU's, RISC speed is not out of the reach of CISC computers. The 060 will be a RISC, superscaler processor. As such, the performance difference between it and superscaler RISC processors may be in favor of the 060. It will definitly be speed competitive with the PowerPC chip. Gary Dunlap : Wyrd Software : It just works like magic!

Fernando, The cost of a RISC would put the Amiga out of reach for all but a very few. Also, there are very few who really need the power of a RISC chip. Add to that the loss of all current software, or the need for a software Motorola emulator (with it's attendant problems) and I doubt it's a good idea at this time. Don

Your assertion that RISC is the way to go is simply that, and assertion. While I generally favo(u)r RISC over CISC, I think the battle is still being fought.

I don't know, I have a RISC based Sun workstation, and no matter how many upgrades, it is always too slow running X-Windows. I think the future of any platform running a GUI will be in multi-processing, and in fact the new Sparc 10 will be a multi processing machine.

I am not so sure that I can agree with your assumption that RISC is 'better'. If you look at what is happening in the Computer industry, people are starting to run into bus speed and memory access limitations. RISC attempts to get better performance by running smaller operations faster while CISC allows more work to be done in a single operation. As you start breaking the 50Mhz range, chip prices start to soar as well as heat and RF problems begin to have an impact. The only way around this is to start doing more things at once (as we are starting to see on things like the RS6000).

After the introduction of the 68040, I'm inclined to believe that the future belongs to CISC processors with RISC-like characteristics. As for existing technology, I'm highly impressed with the latest Transputer CPU (which is RISC). It has a form of piplining that I've never heard of in a sinlge-chip MPU, and the multitasking/multiprocessing/communications built into the chip seem to cry for an Amiga OS conversion to it, for the next big leap forward; From the Byte article I read on it, it seems to have a sophisticated IPC scheme (packet-based). It almost sounds like the Amiga Kernel implemented in hardware, only an order of magnitude more sophisticated and robust. Such an adaptation would not only measure up with state-of-the-art, but surely put the Amiga on the cutting edge.

I agree with John Toebes' message – you can't just say "RISC is better" without studying the application. RISC machines are good for applications that move a lot of data, but if the application is very compute-intensive (multiplication, logarithms, etc.) a CISC machine would probably be better. It would be much more difficult to do a sine or cosine function on a RISC machine than it would be on a 68040. As for the transputer, there are accelerator boards available for the A2000 that contain transputers. The December 1990 issue of "Amiga Computing" called the transputer accelerator card for the A2000 "the ultimate processor accelerator". But then, I'm biased- I work for SGS-Thomson/Inmos, the manufacturer of the transputer! -Bill.