Vamiga: Demo "911 Rescue (1997) (Limbo)" doesn't work

[Gianmarco72]

Trying some demos (from Amiga's demoscene)..

When start the demo vAmiga stops in this screen:

.. and Paula croaks 😣

Tested on SAE:

it works ... and Paula sings 🎶

[dirkwhoffmann]

Great catch! This might be a perfect test case for ruling out some of the remaining bugs (perfect in the sense that the bug shows up very quickly and SAE works well with weakest compatibility settings).

[Gianmarco72]

I think it's a good idea to try some demos too, which often take advantage of custom chips and M68000 (maybe).

[dirkwhoffmann]

This issue is getting interesting, because it's related to the TOD bug of the CIAs. If TOD bug emulation is disabled, the demo works:

It becomes interesting, because the demo runs on a real A500 without any issues. The A500 has the TOD bug which means that TOD bug emulation is wrong.

[dirkwhoffmann]

I think it's a good idea to try some demos too

Definitely. Demos usually push the machine to the limits. Maybe not this particular one. This is the biggest 🤢 I've ever seen 😂. But it serves the purpose. It exhibits a bug.

[Gianmarco72]

It becomes interesting, because the demo runs on a real A500 without any issues. The A500 has the TOD bug which means that TOD bug emulation is wrong.

In other words, the "TOD bug" have a bug.😅

[Gianmarco72]

This issue is getting interesting, because it's related to the TOD bug of the CIAs. If TOD bug emulation is disabled, the demo works:

Found another DEMO that seems to work with TOD bug disabled (and that works fine on SAE):
"5 to 12 (1993) (Syndacate)" (required 1 Mb of chip ram)

But I'm not sure because every time I reset the vAmiga with the button, the audio works randomly. Perhaps it's not related to the TOD bug?...🤔

Furthermore ... after about 3 minutes there's an effect called "Dot World":


.. the audio slows down a few times (but not in SAE).

Maybe a new issue should be opened...?

[dirkwhoffmann]

Found it. It had nothing to do with TOD bug emulation (disabling it simply covered up the real bug). There was an issue in the TOD reset routine. I had to replace

void
TOD::_reset(bool hard)
{
    RESET_SNAPSHOT_ITEMS
    stopped = true;
    matching = true;
}

by

void
TOD::_reset(bool hard)
{
    RESET_SNAPSHOT_ITEMS
    stopped = true;
    matching = false;
}

[dirkwhoffmann]

I'm reopening this issue, because the applied fix breaks #388 and is likely wrong.

It becomes interesting, because the demo runs on a real A500 without any issues. The A500 has the TOD bug which means that TOD bug emulation is wrong.

I've just run the demo on my A500 again, and it failed (it failed both when the disk was inserted on boot and it also failed when the disk got inserted after the hand & disk logo had shown up). Since this contradicts my own statement above, some more investigation ist needed. It might well be that I did run the demo on my A500+ and not on the A500 8A.

TODO: Unpack the A500+ and run the demo there.

[dirkwhoffmann]

Just double-checked that the demo really fails on the original A500 (it does). Hence, I think this issue can be considered resolved.