Please DONT report TF534 Issues here

[8 Bit Dreams]

EDIT BY TF

A forum is the wrong place to report bugs. I was a bit uncomfortable with this forum topic when i saw it but i thought it would be heavy handed to intervene. I've cleaned up the topic now and its been restored because it has some value in it but if you have a bug to report do it on github

https://github.com/terriblefire/tf534

But please remember I am not going to fundamentally change this card for you. The target machine is an A500/+ if you get mileage on a different machine thats great but i am not adding features to disable/enable anything on the board. It is what it is.

---------------------------------------------------------------------------------------------
Hello,
built several cards this weekend and stumbled on strange behavoir:
Cards tested for several hours - no isuues, working flawlessly, however, if i start Shell and type: CPU FASTROM, then start Sysinfo and try to test RAM - Amiga freezes!?
Can someone check on its system and report please?

[FluffyRedLobster]

Hi 8 Bit,

My 534 build is still in progress but I will definitely check this when it's complete (just got a solder issue with one of the ram chips to solve I think). I had somewhat similar results with my 530 (systest would freeze with cpu fastrom, fine without it - there's a thread in the 530 forum that describes the symptoms if you want to compare. Not sure I ever tried sysinfo, just systest / amigatestkit).

Did you see this behavior on all the cards you built?

[terriblefire]

I've never seen this issue. The TF534 hasnt been on my radar much lately.

[dalek]

I can also sometimes reproduce this problem on my card (rev 2c running tf53x_2019_02_18_alpha firmware at 40MHz, no FPU).

There is general flakiness in WHDLoad games, even though Workbench is stable. Things like Frontier run fine, but F/A 18 interceptor does not :/

MBRTest-2 (Microbotics RAM test) and Keir's AmigaTestKit ram test both often freeze the system when run from Workbench with CPU FASTROM.

The DiagROM RAM test runs fine, also if I run the AmigaTestKit RAM test from floppy (no fastrom) it never freezes.

[terriblefire]

Try CPU FASTROM NOBURST

[8 Bit Dreams]

Try CPU FASTROM NOBURST

Turns out this is not TF534 related, Steve.
Have tried and got same results on Matzes 68030TK card.. Well.. Seems, that there is a bug, and CPU FASTROM was removed from 3.1.4 rom for a good reason..
We just can install Mulibs and use mufastrom feature instread of CPU FASTROM.

[dalek]

Interesting - I'll try and test today - I found this at the bottom of the mulibs readme:

The MMU tables generated by the "CPU FastROM" command, an official CBM tool,
are simply wrong if run on a 030 processor. Chip memory is marked as
"cacheable", which is plain wrong. Already spoke to Michael Sinz who agrees
in that point. Don't use it, run "MuFastRom" instead.

[dalek]

Okay, I think with MMULibs this issue is resolved for me. The remaining ones don't seem to be TF534 related:

AmigaTestKit RAM test - I think this fails because CPU FASTROM maps kickstart to fast ram. The ram test then overwrites kickstart memory. Hangs in WinUAE the same way. I should have realised this!

F/A 18 - no idea on this one, but the WHDLoad slave dies the same way in WinUAE A500 with 030 and 4MB RAM (but works fine on my TF330).

[chue]

F/A 18 - no idea on this one, but the WHDLoad slave dies the same way in WinUAE A500 with 030 and 4MB RAM (but works fine on my TF330).

Just putting more info out there - I'm using an EC030, and FA18 under WHDload runs for me. It does die if I try running the demo (first option in the game), but everything else works.

[FluffyRedLobster]

Bit more info in case anyone stumbles across this in the future.

I just finished my 534 build (yay!) and found F/A 18 would bomb out 100% of the time if I choose the free flight option and take off location 4 (the carrier). I can't recall the exact error message now but the memory location was in the $40000000 range. Depending on whether I installed mmulib and enabled mufastrom would change the error message, but the crash was always at the same point. I'd also get occasional crashes with other modes.

This is reproducible in WinUAE if I set 4MB ZIII ram so it's not a 534 issue. If I switch the ram in WinUAE to ZII instead it runs fine.

So back int the real world, I also have a Boobip 2MB trapdoor expansion in my 500 which supplies 1.5MB slow ram and 0.5Mb chip. So I specified the EXPLOCAL option on the whdload slave, which gets the slave to use the slow ram rather than the fast on the 534 - it now runs nicely (albeit at more or less A500 speed).

I'm not an expert, but I guess the slave doesn't like it's ram being in the ZIII space...?