Master MOS disassembly

[james]

This is excellent! I've been able to roll a new 3.20B and 3.50B based on your work that suits my machine's config (though only tested under emulation so far) and leaves a spare ROM slot. Thank you!

[tom_seddon]

Finally got round to testing this on real hardware. It needs more than 1 minute of testing, but it does boot, and my relocation bitmap tool seems to have shuffled things around correctly. (This setup has the relocation bitmaps for BASIC, EDIT, and The BASIC Editor in the DFS ROM.) If this seems to be behaving itself well enough once it's had a bit more testing, I'll make a separate thread about the new ROMs.

Haven't got a disk drive set up right now, but ADFS and DFS seem to work in b2, so hopefully nothing obviously wrong there...

I'll post some notes in a bit about the relocation bitmap tool.

--Tom

[Ukwebb]

I've been using this on real hardware (inc a proper copro not the pi version) and using the DFS with hibasic table etc, and it works a treat imo, and it's noticeablely faster in booting to hibasic than the original version, I mean, we're only talking half a second rather than a full second, but it's definitely faster

[tom_seddon]

I've been using this on real hardware (inc a proper copro not the pi version) and using the DFS with hibasic table etc, and it works a treat imo, and it's noticeablely faster in booting to hibasic than the original version, I mean, we're only talking half a second rather than a full second, but it's definitely faster

Ah, excellent, maybe I can make that thread sooner rather than later then.

I agree about the language transfer speed - it's a big improvement. You can compare the two speeds by timing *HIBASIC (which relocates) or *BASIC (which skips the relocation part), along these lines:

Code: Select all

*KEY 0 TIME=0|M*HIBASIC|MP.TIME|M
*KEY 1 TIME=0|M*BASIC|MP.TIME|M

Timings on Master Turbo:

- MOS 3.20 - 0.21 seconds non-relocating
- MOS 3.50 - 0.6 seconds relocating, 0.41 seconds non-relocating
- MOS 3.50B - 0.38 seconds relocating, 0.23 seconds non-relocating

Those 0.6 seconds feel like an absolute age when you're waiting for them.

I've got a few ideas for further reducing the 0.38 figure. I don't know how much of a difference they'll actually end up making but 0.38 seconds is still a little too long, and maybe it only needs a small reduction to push it into the realms of fast enough.

--Tom

[james]

I’ll add my thanks. I have your release of 3.20 and 3.50 working smoothly in my Master + ‘Turbo’ CoPro.

[tom_seddon]

I’ll add my thanks. I have your release of 3.20 and 3.50 working smoothly in my Master + ‘Turbo’ CoPro.

Good to hear - thanks for the note!

--Tom

[tom_seddon]

I've added some notes about doing a DIY version of MOS 3.50: https://github.com/tom-seddon/acorn_mos ... .50.DIY.md - this is basically me going through the process of sorting out the setup shown in the screenshot above. I figured maybe having a somewhat worked example might be useful. The only particularly fiddly bit is sorting out the relocation bitmaps.

(At the moment, the GitHub releases don't include the relocation bitmaps, so you need to be set up to build the ROMs locally.)

--Tom

[tom_seddon]

I made a separate thread for the new versions: https://www.stardot.org.uk/forums/viewtopic.php?t=28879

--Tom