Individual Computers - New accelerator cards for A600 & A1200

[Retrofan]

The other day I posted something similar, but later I deleted it.

With WinUAE in my Samsung Tv using a Pc I can't boot my system having the latest ACATune enabled... But I can boot perfectly if I use my laptop with the same s-s, without disabling anything.
Another thing: In WinUAE as mfilos says it shows the Acatune -status, and another, if you disable it in the miggy it will report anyway the Acatune -status.

[mfilos]

I just found out why Retrofan...

Possibly in your laptop... you HAVEN'T enabled speccific chipset but left Generic!!!

If you leave it like this:



...then ACATune reports that it doesn't find an ACA accelerator!
If you choose A600 then you get the ACA630 or A1200 and you get ACA1230!!!

[Retrofan]

Surely that's right mate. Anyway at least it detects (or makes Acatune work) my "undetectable" ACA using Amikit in the miggy.

[Schoenfeld]

It's correct that we now check for presence of a certain Denise/Lisa ID in order to make a difference between A600 and A1200 accelerators (that was the routine that was buggy before, this chipset-recognition is new in V1.7).

However, ACAtune is meant for computers with an accelerator. An emulator cannot have any ACA card. I find it a bit far-fetched that a piece of software that is specifically made for a special hardware is expected to work on an emulator which does NOT emulate any of the ACA accelerators.

Toni Wilen has the full documentation of the ACA620, which also includes part of the ACA card recognition. He may be able to fix UAE in a way that it reports "no accelerator found". I refuse to add complexity to ACAtune just for compatibility with a weird case. It's like adapting a car tire for use on a bicycle.

Jens

[Lord Aga]

Really now, why would anyone run ACAtune on WinUAE

[Toni Wilen]

If it can happen on real hardware using some other (possible non-existing) accelerator: somebody else's problem.
If it can't happen (wrong address decoding/whatever on unused address space etc..): will be fixed.

Hardware specific software rarely works as expected on other hardware. ACAtune previously "working" was probably only accidental.

[mfilos]

@Schoenfeld
I never said that I was expecting you to add complexity to ACATune just for such a case (...although a car tire in a bike was really sexy as the Batpod... )
Just realized the problem and thought about commenting it for other people to know. Since I can workaround the problem using 2 options, it's ok for me

@Lord_AGA
Apparently you haven't read my #2600 post...
I have an A600 that runs just fine. Occasionally I backup it's configuration to WinUAE to be able to check other things and then transfer them back to the Amiga (via CF card).

Apparently ACATune doesn't run on WinUAE (and that's why till 1.7 was saying that it didn't find an accelerator and all was good) since custom kick that I made was being used via the ROM choice in WinUAE.
Now I have to disable either the speciffic chipset (which is an option) or disable ACATune (which is an option as well).

@Toni
Thanks for taking a look at it.
If I can help with any way just gimme a shout

[Retrofan]

The only problem I find is that for example I wanted to see if there was still a problem with the Executive program, and as the ACATune -status was working the same I thought that then there wasnt any problem now with it.
I saw with the other ACATunes that Executive disables it. Now I just have to suposse that it has to be a problem for it yet, as ACATune -status always will show it working now.

[Lord Aga]

Originally Posted by mfilos

@Schoenfeld
@Lord_AGA
Apparently you haven't read my #2600 post...
I have an A600 that runs just fine. Occasionally I backup it's configuration to WinUAE to be able to check other things and then transfer them back to the Amiga (via CF card).

Apparently ACATune doesn't run on WinUAE (and that's why till 1.7 was saying that it didn't find an accelerator and all was good) since custom kick that I made was being used via the ROM choice in WinUAE.
Now I have to disable either the speciffic chipset (which is an option) or disable ACATune (which is an option as well).

Ah, ok. I can see how that's useful. Still, I wouldn't expect WinUAE to support such a specific hardware + software combination. If it can be done somehow - great

[MrDBUG]

- Just wanted to ask if the ACA1230/28 is supposed to work as intended now, mine has been sitting idle in the box as I lack the skills to do timing fixes or dare to try another go at the installation after getting card back from repair last year. :-/

Tempted to just sell it to someone who can use it .. ! :-(

- Or ship it to AmigaKIT to install in one of those Petro A1200s if they are getting any .. (? I dont know if hes just going to sell them by himself .. ?)
Hope for the best, I guess time will tell.

[Schoenfeld]

The ACA1230/28 will work after you've done the board fixes. The newer cards (ACA1231, ACA1220, ACA1232) are slightly more tolerant, but they still don't work with all the assembly errors that Commodore and Escom have left us with.

The most common fix is also the easiest to do (remove those two caps). You should find someone locally to have it done, so you can enjoy 64MBytes of memory and some speedup.

Jens

[MrDBUG]

( On a lighter and offtopic note :-D The Chameleon is great :-D
Just put in the Beta8 Core tonight. )

[vroom6sri]

MrDBug - I am not all that skilled but the A1200 is quite easy to disassemble and reassemble. I took mine apart easily enough so I was just left with what looks a bit like a PC motherboard and then took it to a local TV repair place who charged me £10 to remove the capacitors I had identified thanks to the information someone in this (and probably other) thread(s). It was money well spent. Having a "professional" do it meant the job was done neatly and gave me peace of mind but more importantly it worked! I sold my ACA1230/28 on so I could get the ACA1231/42 which is really good but the 28 is still a great card.

[demolition]

Originally Posted by mfilos

Apparently ACATune doesn't run on WinUAE (and that's why till 1.7 was saying that it didn't find an accelerator and all was good) since custom kick that I made was being used via the ROM choice in WinUAE.
Now I have to disable either the speciffic chipset (which is an option) or disable ACATune (which is an option as well).

Apparently this is not just an issue with WinUAE as Acatune detects an ACA630 in my A600 when there's no accelerator in it (only +1MB chipmem was installed)..

[mfilos]

Aye, that's true demolition mate.

Apparently the logic in which Jens and his team worked on the new version (to avoid the issues with accelerators NOT being recognized on some revisions) was that:
- If someone runs ACATune, apparently runs it because owns an ACA6x0 or ACA12xx
- If Chipset is AGA then ACATune returns that the board is an ACA1230@28MHz (even if an ACA12xx isn't physically installed)
- If Chipset is ECS then ACATune returns that the board is an ACA630@25MHz (even if an ACA6x0 isn't physically installed)

Ofc if you have correctly installed an ACA accelerator and you don't have a problematic revision or case, then your accelerator will be recognized just fine in it's correct clockage.
If not... then it will get recognized with the generic identification that I wrote few lines back.

I kinda dislike the logic imho, but I guess it's really difficult to debug and troubleshoot EVERY single revision and incident that some mobos have :S

[Photon]

Which is the latest stable ACATune version for use with my ACA630@25? I have a version which gurus or "suspend reboot" (sometimes it works) when turning on MapROM.
Edit: never mind, 1.7 IDs my card correctly, seems to work fine with maprom, and -vbrmove worked.

I would also like to know what settings special to ACA (FastChip etc), if any, which are incompatible with "blit to chipmem+CPU reads or modifies result" (ie. which options could cause chipmem to be cached) and "CPU decrunches code, then jumps to code" (ie. lines in the ICACHE that were in the decrunch area don't match lines to the same address in DCACHE).

[fitzsteve]

Originally Posted by Photon

Which is the latest stable ACATune version for use with my ACA630@25? (I will try 1.7, but other suggestions are welcome.) I have a version which gurus or "suspend reboot" (sometimes it works) when turning on MapROM.

If someone knows, I would also know what settings special to ACA (FastChip etc), if any, which are incompatible with "blit to chipmem+CPU reads or modifies result" (ie. which options could cause chipmem to be cached) and "CPU decrunches code, then jumps to code" (ie. lines in the ICACHE that were in the decrunch area don't match lines to the same address in DCACHE).

I'm using V1.7 and all I have right now is Maprom & Maxmem settings. I need to do more testing. A list of what commands work with the ACA620 would be useful, they're not printed in the manual

You can get ACA Tune V1.7 here:

http://eab.abime.net/showpost.php?p=...&postcount=239

[Retrofan]

Yes, V1.7 is running fine; as fine and fast as the olders. It only can give you a problem if you take your Cf to use with WinUAE, where it's better to disable it in the s-s. . I'm using it because I use a "rare" Workbench (BTW I'm improving it), where it also runs fine (Thanks again, Jens).

[mfilos]

Version 1.7 is basically for new cards (ACA620, ACA1220, ACA1232) and also has a new logic (to avoid the issues with accelerators NOT being recognized on some revisions. More on #2615).
Since you use an ACA630, using the old 1.5 is same as new 1.7

[Photon]

Originally Posted by mfilos

Version 1.7 is basically for new cards (ACA620, ACA1220, ACA1232) and also has a new logic (to avoid the issues with accelerators NOT being recognized on some revisions. More on #2615).
Since you use an ACA630, using the old 1.5 is same as new 1.7

Cool, thanks for quick answers See edit above... unless someone recommends otherwise it's OK to use 1.7?