memtest cart for SAMS/Myarc/Foundation/Corcomp/etc

[+jedimatt42]

I've been kicking around this memory test tool for quite a while. Just updated it to include Corcomp 256k/512k testing... 

 

https://github.com/jedimatt42/ti994a-memtest/releases - version 2.2 now. (release is tagged 2.2_a cause I made a mistake) 

 

With a Corcomp board at crubase >1000 my old memtest messed up the cru bits while checking for the presence of the myarc card... this solves that, and adds Corcomp size detection and specific paging support. 

[+jedimatt42]

Oh, some will note, this used to be at a '32k' specific url... I renamed that git repository since it was misleading. 

[+Ksarul]

I use your earlier version regularly when testing memory devices. I have found one oddity though: sometimes, a 4M SAMS board is recognized as a 64K Foundation board for some reason.

[+jedimatt42]

2 hours ago, Ksarul said:

I use your earlier version regularly when testing memory devices. I have found one oddity though: sometimes, a 4M SAMS board is recognized as a 64K Foundation board for some reason.

They (Foundation and SAMS) use the same crubase to control paging.  A fault in SAMS page count assesment can cause this... If the SAMS doesn't appear to have at least 128k then I reject it as SAMS and try other board types... I'm not looking at DSRs, just crubit behaviors... looking for a memtype by precedence:

 

1st, is there mem?

2nd, try SAMS paging, compute page count, if at least 33 pages, then it's sams... this probably improperly excludes 128k AMS, but those can just sit in a dungeon somewhere.

3rd, try Corcomp... 2 crubases to check, and must be able to turn ram off completely... 

4th, check for Myarc at >1000 

5th, check for Foundation at >1E00 <-- faulty SAMS can look like this...

6th, assume it is one of the many standard 32k boards.

 

It is written that early foundation cards don't have DSRs, and it is known that many foundation cards have Myarc ROMs... so checking for the DSR is also not definitive. 

 

I could tighten this up though... a foundation card would actually have 4 or 16 pages (if hand upgraded to 512k), never 2. If anything has only 2 pages, it is just an error detecting and I should stop right there. 

[globeron]

Have not tried it yet. But does this work for Horizon 3000 ramdisk and Rambo extension on it  to test memory?

[+jedimatt42]

3 hours ago, globeron said:

Have not tried it yet. But does this work for Horizon 3000 ramdisk and Rambo extension on it  to test memory?

No. It is focused on the things that page in the 32k expansion space. Thus it runs itself from ROM space.

 

Mostly cause prior tests I could find would 'pass' when I knew there was a fault. 

 

Supported cards: https://github.com/jedimatt42/ti994a-memtest#supported-cards

 

I no longer have an HRD, but I assume with the recent re-release of HRD-4000s and it not conflicting with test code living in 32k space, there must be a trusted test out there already. Any other reality seems foolish.

 

Note: RAMDISK testing in my program destroys any data in the RAMDISK. 

[globeron]

Thank you for the tool. Seems the Corcomp 512K is all okay.

 

Foe EXP-RAM TEST first run got an error already at page 0. Second run all okay until page 15.

 

For EXP-RAM BURNIN, it seems to be in an end less process? Going to differerent pages like 95 as well. After an hour  I stopped at Burning Pass 12 

[+jedimatt42]

Yep. Burn in is endless... But page on a 512k corcomp should never be more than 15 (0-15 32k pages).  So there's a bug I need to fix.

 

The corcomp card I tested on is also subject to intermittent failure. If it is rare like that, might just be edge card contacts.  

 

[+dhe]

Does someone have a pre-built  exptest_c.bin that I can download?

[+jedimatt42]

4 hours ago, dhe said:

Does someone have a pre-built  exptest_c.bin that I can download?

https://github.com/jedimatt42/ti994a-memtest/releases

 

Click the triangle to the left of the word 'assets', and it will expand to show the .bin