Recent content by Argyle Gargoyle

Yea it works on rp2350 - see details here: https://github.com/evansm7/pico-mac/issues/7 Sadly, no sound yet…

Those bus drivers getting hot is a clue.You should look for a short on the lines that the hot 651’s were trying to drive. You could use a DMM to check for continuity to ground on those bus outputs as a first check, although they wouldn’t necessarily have to be shorted to ground. When you power...

This is really exciting - congrats Snial on a big landmark!

Snial, does everything run in a single thread, or are there useful speedups to be had with using two cores? Like a “read ahead” of the code to prefetch stuff? Perhaps it’s a question for one level higher, eg pico-mac or equivalent.

Those tricks are fun. If you haven’t ever seen the book “Left Brains for the Right Stuff: Computers, Space, and History” by Hugh Blair-Smith, it’s a fun read about the processor designs leading up to the Apollo project. Hey, it looks like rp2350 has hardware udiv and sdiv- could stuff like that...

I think you can set up a DMA to fill the PIO fifo at full speed, and then set the PIO to run at 15kHz. The DMA can be paced by whether the fifo has space, so you could initiate it and it would keep feeding the PIO as needed. Which is pretty much exactly what you said, except that “copying the...

We need a countdown! What is it, about 25 instructions left?

Really interesting how the cost function changes when your host clock is 20x faster than the emulated cpu. Also when the number of regs differs. You mentioned earlier that you use global registers, but I’m not sure I understood - does that mean that A0-7 and D0-7 each get stored in a (cortex)...

Maybe it’s too early to ask, but do you envision the emulator core following the same conventions/interface functions as e.g. musashi? So a pico-umac build could swap in MØBius as the engine? I wonder what other platforms it might also work with? I guess mame in its full form is out of scope...

Nice! So going back to the testing/validation question: how do you structure a test of the emulator core? Do you emulate the Cortex code so you don’t have to run it on pico hw? I’d imagine there are specific tests to run, but after the emulator is passing those, would you do something like...

That’s really a brilliant idea - it was the rewriting QuickDraw that was sounding like a lot of work. Aside from enabling a larger screen, you’d eventually be able to use the rp2350 hstx to do DVI But I agree with your sentiment of getting the simplest thing working first.

Yea, it sounds kinda complicated. I’ve also been looking at the rp2350 a bit more, and it supports external PSRAM (up to 32MB). There is some throughput info here: https://forums.raspberrypi.com/viewtopic.php?t=386630 I wonder what the speed hit would be if you tried to use PSRAM, and if you...

If you build it, they will come! Hey, speaking of ram usage: how feasible would it be to reclaim the ram ordinarily allocated for the screen? If the mac uses those 640x432 pixels as an output only (ie never reads it back), couldn’t we pipe that data to a second rp2040 that would be dedicated to...

I think it’s interesting because I think the pico is a much better platform than a raspberry pi since you don’t have to wait for Linux to boot - it feels more like real hardware! Plus, the newer RPis seem to have gotten pretty far from the $5-$10 price point, but pi picos are still really cheap...

Hey Snial, weren’t you working on an emulator of your own? How far along is that?