What do people think of the idea of a DSP upgrade for the Falcon? In terms of difficulty, expense, and benefit.
Aside from people overclocking their DSP's (up to 64mhz) it hasn't really had the same attention over the years as what the CPU has!
It seems the DSP 56k family
didn't stop development until only a few years ago, still supported and manufactured, so there are some modern 56001-compatible chips unlike what we have with the 68k family. Eg the DSP 56724
- dual core, 250mhz (500 mips), US$6.21 in quantities of 1000.
Not much software uses the DSP, and we know it wouldn't make the machine any faster to use day-to-day, but it should be a pretty low effort upgrade too, shouldn't it? (I'm not a hardware guy, so I'm not sure). At $6.21 for the chip, a few dollars for a few MB of SRAM, and not needing a whole lot of logic to integrate with the rest of the Falcon chipset (I assume? the DSP is pretty stand-alone, aside from DMA access?) the board would be the most expensive part. It seems to me it would be the cheapest, easiest, and most "true to spirit" of the original Falcon to get 500 mips of power available for video decompression, or things like BadMood and Quake 2 if anyone like DML were interested in playing with it.
Failing an upgrade of the Falcon, it also seems like it would be a good add on for the CT60. At 500 mips, it would have a similar relationship with a 100mhz '060 to what the old 32mhz 56001 was to the 16mhz '030. (Currently, the '060 is so powerful it's not worth bothering with using the DSP, but a 500 MIPS DSP added to the CT60 might change that!)
Anyway, don't mean to annoy anyone, from reading hardware threads started by non-hardware people over the years I would stress that I DON'T expect anyone to make this, nor am I even convinced it's a good idea :p Just an idea I thought it would be interesting to see discussed!