DSTB1 exxos first tests & experimental firmware.

[exxos]

Interesting twist.. GB6 will run fine, along with YAARTTT etc. but when frontbench loads it crashes not long after it loads :( the second test just came up with 2 bombs and black screen with nothing on other than the frame counter. So not sure what is going on there.

IMG_0521.JPG

EDIT:

Tried EMUTOS It locked up after 6 six frames and the sound is like stuttering playing in the background.

REMIX! :lol:

[Badwolf]

REMIX! :lol:

Funky.

But, yeah. Timing gets out of whack when trying to serve two masters. :(

BW

[Badwolf]

So @exxos, I didn't have any problem with yaartt or frontbench, but MiNT won't start up: bus errors starting the desktop in the altram range.

I think this is what I found before: MiNT, I suspect, puts a lot more through the SDRAM controller and it can't keep up with that extra AND term.

I think perhaps a hardware solution is needed (or a better SDRAM controller). I was thinking about a diode between AS and AS_INT or similar: let blitter pull down the AS_INT line [so AltRAM can trigger], but AS has to be pulled down by the logic.

Either that or bridge AS and AS_INT, remove one from the CPLD logic and just let AltRAM accesses propagate out over the motherboard.

Haven't experimented either way yet.

BW

[exxos]

@Badwolf odd you can run FB and I can't. Maybe if you sort the standalone code I can plug it in with my 16mhz booster to see what happens ?

[Badwolf]

@Badwolf odd you can run FB and I can't. Maybe if you sort the standalone code I can plug it in with my 16mhz booster to see what happens ?

Yep, OK. I'll be coding this one 'blind' as can't test it myself, but you're actually half way there with the blitter mods. Mostly it'll be removing the BGK requirement.

Will have quick nose this arvo if I can.

BW

[exxos]

BW sent me some updated code to try the RAM problems out with. This setup runs in "slave mode". Where the CPU is in the H5, not on the booster. So no 16MHz boost on the CPU in this setup.

Currently it's starting its second RAM test loop. It failed after a couple of tests before. So at least this setup rules something out!

IMG_0532.JPG

[exxos]

Ram passed fine for a couple of hours. But altrom.prg and GB6 won't run in alt-ram :shrug:

Seems to corrupt the floppy contents.

IMG_0536.JPG

GB6 loads from floppy fine in STRAM. I re-ran YAARTTT and it's passing. So no idea what's going on there.

IMG_0537.JPG

[Badwolf]

I should just, for clarificaiton, point out that everything bar the first few posts in this thread are using experimental firmwares to try out different options and to try to get the board working in it's secondary 'AltRAM slave' mode.

Please note the stock board with the stock firmware works fine with any already TOS2.06-enabled machine, or with any TOS1 machine using EMUTOS.PRG.

BW

[Badwolf]

Developing firmware for AltRAM slave mode: no piggy backed CPU and plugged into an H5 expansion port

Ram passed fine for a couple of hours. But altrom.prg and GB6 won't run in alt-ram :shrug:

Seems to corrupt the floppy contents.

GB6 loads from floppy fine in STRAM. I re-ran YAARTTT and it's passing. So no idea what's going on there.

OK, a couple of things this could be.

1) YAARTTT does all its accesses in word-sized chunks. Possible byte-level access is screwed, although that part of the code shouldn't have chanegd.

2) YAARTTT is actually 'quite slow' between RAM test accesses. A program loading and running from AltRAM will be hitting up different addresses back to back. Ie. accessing it quite a bit faster. It possible the timing is off.

Point 2 seems more likely to me ATM -- it's been built and tested with a 16MHz chip -- but I went to lengths to try to make sure it behaved at other reasonable speeds. Possible I've messed something up there, of course.

Would need to have a bit of a think about it.

Point 1) should be verifiable with byte-level peeks and pokes in STOS, BTW, but I'd be surprised if it were that.

BW

EDIT:

Quick thought: you are running ALTRAM.PRG from the auto rather than the desktop, yes?

[exxos]

should be verifiable with byte-level peeks and pokes in STOS, BTW, but I'd be surprised if it were that.

Will try it.

Quick thought: you are running ALTRAM.PRG from the auto rather than the desktop, yes?

No. I just tried from AUTO and the desktop gets more mashed. "TEMP" isn't a file on the floppy even.

IMG_0538.JPG

Might rename this thread and start again with your firmware in another thread as like you state, its getting confusing.