A query on "Errors in memory testing" (ACA 1234 + ACA 500 Plus)

Great, you go and make it ‘easy’ after the NOOB! (Me) uses it.

Quote from ljmarent

notes:

I built my 4GB flash CF install myself..using the ACA500+, OS3.2 install files, a bunch of update files from iComp and and AmiNet downloads.

The TF536 is new to the house and should leave soon, I don't think anything in my software config is more friendly to the other accelerator. I've made zero changes to help or assist it. It's running the exact same CF card and software load as my ACA500+

Not completely "the same" - please try to deactivate a few of the ACA1234 features, such as PCMCIA CF card driver, mounting of the Install disk and maybe even the local CF slot. Maybe it's a software conflict with the software that the ACA1234 starts from it's flash.

Quote from Jens

Not completely "the same"

Well ,,, from my perspective it’s exactly the same.
I take the CF card out that I built for my Checkmate ACA500+ & ACA1234 rig and put it into the TF536 rig… it’s the same card… untouched…

Perhaps I don’t understand what you mean.

How do I deactivate ACA1234 features ?

Last time I ran the ACAtool for this card, all I could change was clock speed, everything else was greyed out ?

and does the ACA1234 load software ?

I thought only the ACA500+ was doing that and the ACA1234 was just providing CPU and DRAM in my setup … but now I’m totally interested in what is happening between these two iComp cards in my setup.

Quote from ljmarent

Last time I ran the ACAtool for this card, all I could change was clock speed, everything else was greyed out ?

You do have the possibility to deactivate mounting of the install disk.

Quote from ljmarent

and does the ACA1234 load software ?

Sure it does - from it's own flash. The driver for the local CF card slot is loaded during the autoconfig phase, which happens right when you leave the ACA500plus menu.

Setting jumper JP0 will not only reset all settings, but also keep the autoconfig process from loading the CF card driver. Who knows, maybe WHDload doesn't like something of that? Note that JP0 will also slow down the card to a safe 25MHz.

Okay, so you are suggesting I put the JP0 in the 25 MHz safe position and test again ?

Quote from ljmarent

Okay, so you are suggesting I put the JP0 in the 25 MHz safe position and test again ?

Yes, please do that.

Okay..some good news

With the JP0 jumper in the 25MHz Safe position (edge of card position)
Supremacy and Unreal now work without causing errors in the first 60 seconds

I can actually play the games YAY!

Agony does not crash, but the game doesn't work right and the graphics once you get to the start...
Oh...let me check.. I was tinkering with that games tool types ...let me reset that one..
I'll come back in a moment...

Okay...

I reset WHDLoad.prefs to stock settings and rebooted

Agony no longer crashes. but it's not playable either

It's works fine with just the ACA500+

But once you put the ACA1234 (safe mode) into the system there is all sorts of weirdness...

Also... Xenophobe is behaving like Agony ... the intro and logo's work, but once the 'game' starts the video is unusable until you die and then the menus come back and it works.. Maybe there will be WHDLoad settings that sort that..

anyway..there is a definite interaction between the ACA1234 firmware and WHDLoad as well as Amiga Test Kit
Disabling the ACA1234 firmware stops the crazy error messages and invalid instruction stuff

Will the setting of JP0 remove any Kickstart-ROM out of the ACA1234-flash-memory?

Quote from ljmarent

Disabling the ACA1234 firmware stops the crazy error messages and invalid instruction stuff

OK, that is something we can investigate. However, a better explanation would help:

Quote from ljmarent

But once you put the ACA1234 (safe mode) into the system there is all sorts of weirdness...

That's nothing we can try to debug - "weirdness" is nothing that tells a programmer where to look. If there is something showing on screen, please describe it, or take a screen shot. If it's a crash, please write down the Guru number.

Quote from OMo1972

Will the setting of JP0 remove any Kickstart-ROM out of the ACA1234-flash-memory?

It will not alter the flash contents - only the EEPROM contents to "factory setting", and it'll keep the computer from executing the flash contents.

Does it make sense to try the JP0-method before updating the CPLD?

Quote from OMo1972

Does it make sense to try the JP0-method before updating the CPLD?

Not at all, as this will keep the autoconfig software from running, and it'll keep the offending 1MByte block available for any software. Only if you block the memory with a separate tool (see earlier posts in this thread), this may help.

Quote from ljmarent

Disabling the ACA1234 firmware stops the crazy error messages and invalid instruction stuff

Quote from Jens

OK, that is something we can investigate. However, a better explanation would help:

FYI, I luckily can't confirm those symptoms on my setup (A1200, ACA1234, plain OS3.1). I have activated all the features on the ACA1234 and could play the WHDLoad installs of Agony, Supremacy and Unreal without any hassle (Supremacy still plays nice nowadays, the other two, well...).

The problems described from ljmarent seem to depend either on the combination with the ACA500+ or on a software conflict, not on the features "per se"?

"I have started with the ACA500 plus connected in conjunction with the ACA 1234."

Just a thought..

But the ACA500+ & ACA1234 combo is still causing a lot of software to not run properly, especially when the CPLD patch is not applied and the ACA1234 is not in safe mode.

If you start with these two cards connected and try to perform the ACA500+ firmware update...

I'd be worried the updater might also be affected by the aforementioned issues...

I'd certainly not recommend updating the ACA500+ with the ACA1234 installed, until iComp has tested that specific config and found it safe

Quote from ljmarent

I'd certainly not recommend updating the ACA500+ with the ACA1234 installed, until iComp has tested that specific config and found it safe

I have executed the V0.140 update before I even knew that there was a memory mapping bug.

Quote from diegoarevalo

Now when I start the AMIGA 500 the ACA 500 plus does not turn on. The ACA500 plus is totally dead.

That's what the de-brick mode is for: You can bring it back to life by holding down the DisMo button until you see "0_" in the DisMo (the underscore is blinking). Once you see the blinking, do a keyboard reset, and the machine will come up as if the ACA500plus is not connected.

Now you can run the update-ADF from a floppy emulator, or you can write it to an actual floppy and boot that. It will perform the update, and you can then short-press the DisMo button and reset to use the menu again (or power cycle).

Quote from lucadip

I have activated all the features on the ACA1234 and could play the WHDLoad installs of Agony, Supremacy and Unreal without any hassle (Supremacy still plays nice nowadays, the other two, well...)

Hey ... I don't come to your house and tell you what games not to play

But honestly... I'm glad the ACA1234 appears to be working on the A1200 now..
Though... I read lots of people saying 'flawless' before the memory corruption bug was fixed...
So please don't get complacent...test...test...test...

ACA1234 'should' be a super star Accelerator...if everything works as it should.

Quote from lucadip

FYI, I luckily can't confirm those symptoms on my setup (A1200, ACA1234, plain OS3.1). I have activated all the features on the ACA1234 and could play the WHDLoad installs of Agony, Supremacy and Unreal without any hassle (Supremacy still plays nice nowadays, the other two, well...).

The problems described from ljmarent seem to depend either on the combination with the ACA500+ or on a software conflict, not on the features "per se"?

did you patch the CPLD ?

Because that bug still affects 1200 users until patched.

A little report while I'm waiting for the cable to do my CPLD update & doing more tests.

As reported before, I've flashed the firmware hotfix but it didn't help with the diagnostic software (AmigaTestKit & MBRTest-2 still fail/freeze).

Now I tried the "problematic" WHDLoad games reported by ljmarent. Supremacy and Unreal crash on me every time with the ACA500Plus + ACA1234 config and work flawlessly without ACA1234 (or with ACA1221lc in its place). Interestingly, Agony seems to work though.

I also installed the new ACA500Plus menu and tried fiddling with various settings but it had no effect. Hopefully the CPLD update will fix the issues, but I'm a bit suspicious that the firmware hotfix doesn't fix anything at all (at least in my case).

Maybe it's a silly idea: I have 2 MB chip RAM (thanks to rev 8.1A mobo with fatter Agnus and appropriate RAM expansion from icomp) - could it have anything to do with my problems? It's not a "typical A500" config after all.