Quote:
Originally Posted by Gorf
I would have no problem with utilizing a slave linux/bsd kernel on the ARM side for drivers (usb, ethernet, drives, display initialization, etc). It would do it's work in the shadows and a user would not have to care about this.
|
An ARM I/O processor could probably offload some of the 68k processing. The big work would be creating all the software drivers (half on ARM side and half on Amiga side). FPGA experts would probably be needed just to bring the board up with 68k+Amiga chip set and provide a drive of some kind. You would want a lot of Amiga developers onboard with that board configuration to make it worthwhile. The standalone FPGA boards like the FPGA Arcade and MiST already have an FPGA with ARM configuration but don't have as big of an FPGA and cost much more because they are not mass produced. A 2nd generation FPGA Arcade is planned with a bigger and more modern FPGA.
I thought about trying to license the Apollo Core as well. I would need the license to include all HDL sources and perpetual unlimited modification and distribution rights for multiple devices. Perhaps MikeJ would be interested in creating a more practical and compatible 68k core with it if he could include it in his FPGA Arcade and the sources are not too much of a mess. I do like the idea of moving to a large FPGA where there is room for a 68k compatible FPU, at least a 128 bit wide SIMD unit (not sharing the integer register file so floating point could eventually be added), MMU with x86 PAE/ARM LPAE type support and multi-core support to work toward an eventual ASIC. I doubt I would receive as favorable of a license as the Vampire and it may be more cost effective to decap a 68060 and turn it into sythesizable HDL (costly and time consuming but the patents should all be expired).