> On Thu, Nov 4, 2021 at 5:27 AM smf <smf_at_null.net> wrote: >> >> You could make that argument about all of the suggestions. >> >> 1. Adding an extra vic 2 would be a nightmare. The cpu's clock comes is generated by vic2 from the dot clock and now you have two vic2's. They don't have a reset input, so the clocks will be all out of phase. > > The best solution for adding a second VICII would be to use a > dual-port ram. The 8mhz clock wouild still be handy but it wouldn't > be necessary to worry about timing between the two ViCII's. > Having attempted to synchronize two VICII's in the past I learned they > don't reset their rasters at all on powerup, starting at a random > point on screen. :) Ok, ok... external VIC-II card crossed off the list for these reasons and others even I can understand. > >> 2. Digital sound doesn't need a clock as high as the dot clock, the ø 0 clock would be good enough and you would need to synchronise to that anyway. This would work as "send data to an external sound processor with its own memory and let it do the work". The dot clock would allow for internal upsampling or extra DSP effects that the SID wouldn't be able to do and would be too expensive to do in software by the CPU. >> >> 3. The math coprocessor one is a bit vague, you could hang a raspberry pi zero off the back of a c64 but I'm not sure why you'd use the dot clock. The ability of doing DMA from outside would be helpful so the CPU would ask for certain calculations to be done by the external math coprocessor by presetting certain memory locations with both the operations and the data to be processed, and then asking the external chip to do them. Things like 3D games, database applications, spreadsheets, graphics suites... would benefit from it. > > Perhaps this would be similar to the 68881 add-on for the apple2 > series, which did actually accelerate math functions that called thes > SANE toolkit in GS/OS. > >> On 03/11/2021 23:09, gsteemso wrote: >> >> As much as I like that idea, I'm not sure how it would take advantage of a pixel-rate clock signal. >> >> On Nov 3, 2021, at 11:59 AM, Justin <shadow_at_darksideresearch.com> wrote: >> >> TCP offload. >Received on 2021-11-04 21:00:02
Archive generated by hypermail 2.3.0.