Memory testing.. any idea whats going on here? [message #395339] |
Wed, 03 June 2020 23:25 |
|
Originally posted by: chris.todd
This is on a Rom3 GS with a Transwarp GS running at 10MHz, 32k cache using the SHH Systeme cache card. Memory card is a ByteBoosters DarkRAM 8MB.
Whenever I run a memory test - either BitBanger or the ReactiveMicro MegaMemoryTest I get very odd results.
With MegaMemory test, I get “errors” detected that don’t appear to be errors, for example:
“Bad memory word detected at $27D8C8. Read $0008. Expected value $0008”
“Bad memory word detected at $279358. Read $0040. Expected $0040”
For each “error” the “read” and “expected” are equal - so how is that an error?
BitBanger simply complains that it found a “fatal error in my accelerator cache”.
The TWGS passes all self tests even on continuous loop for a couple of hours, so to me it feels doubtful that it’s a cache issue, especially with a more “modern” cache board in there.
One thing I did notice is when I upgraded my TWGS from 8k to 32k and forgot to cut traces TC4/TC5, everything still worked as expected. The TWGS actually recognized 32k cache and passed self tests and booted just fine.
The system seems relatively stable, so I’m not sure if I should be concerned with these random memory issues in every tester I try.
The TWGS also has a new ReactiveMicro cable, replaced power filter capacitors, and the GS has a brand new Ultimate PSU from Reactive.
Other cards are a Carte Blanche (Original), CFFA2000, Uthernet I and a clone Z80 Softcard.
Should I be concerned about these memory errors that show up on these tests? Any idea why a matched value would show as an error?
|
|
|
|
|