NEWS | PROJECTS | CONTACT | LINKS

NEWS ARCHIVE (OLD to NEW) : 1


06/04/2026 ~ Super Alive and...kicking!

Long time no writing but I never stopped developing the emulator. The biggest change so far is that I completely changed my approach.

The breakthrough didn’t come from new code. It came from changing the way I looked at the system.

Instead of forcing the emulator to follow the header vectors, I let the 68020 follow the actual execution path inside the ROM. Once I stopped correcting the ROM and stopped “protecting” the CPU, the machine behaved exactly as the original hardware does.
Yup , the emulator is now as CD32 as it gets!

The ROM itself pushed me to unify the Kickstart and Extended ROM into a single 1MB image (instead of two separate parts), remove all surgical patches, disable the overlay permanently and synchronize the shadow vectors with the real ones. It’s amazing how clearly it was trying to “speak” once I stopped interfering.

The result was immediate:
No resets, no corrupted vectors, no dead ends.
The CD32 entered the Extended ROM cleanly and executed real initialization code for the first time.

The next step is straightforward now:
bring up the display pipeline and produce the first visible pixels.

I feel the boot menu is closer than ever.

PS - Two weeks ago the site actually crashed from the amount of visitors.
Thank you all for the interest and support.



20/01/2026 ~ Into the Extended ROM?

The breakthrough finally came! Τhe emulator wasn’t crashing, it was signalling!

Once I stopped trusting the header vectors and followed the actual execution flow, everything aligned and stabilzed. The Kickstart trap at 0xF800B6 wasn’t a failure. It was the key. And with a safe SSP, a relocated RAM handler and the overlay permanently disabled, the system finally stepped through the door it had been knocking on for days.

The CD32 has officially entered the Extended ROM.
For real.

The trace shows it clearly:
MOVEM.L, LEA, JSR —> the real initialization code running exactly as Commodore intended.

No loops.
No corrupted vectors.
No dead ends.
Just clean, deterministic execution.

The next milestone is obvious now:
first visible pixels.

Paula32 has officially crossed into a new era.


08/01/2026 ~ The Breakthrough

After weeks of chasing phantom bugs, the real breakthrough came from a completely different angle: the CD32 wasn’t failing , it was waiting!

Once I stopped treating the resets as emulator faults and started treating them as forensic signals, everything clicked. The Extended ROM wasn’t “dead” or “incomplete”. The CDUI wasn’t missing. The RomTags weren’t lying. The system was simply expecting hardware responses that no emulator had ever given it. So I rebuilt the environment from the ground up:

• 7 problematic modules bypassed with surgical precision
• Fake ExecBase to stabilize the bootflow
• Stubbed libraries to keep the firmware alive
• Akiko/CD region mapped and neutralized
• Forensic jumps to keep the execution path honest

And then it happened!

The CD32 crossed the point of no return:
Extended ROM execution, AGA writes from E0xxxx, and full video init logic alive and running.
No more resets.
No more dead ends.
Just pure CD32 firmware doing exactly what Commodore intended!

For the first time, Paula32 isn’t “trying” to boot the CD32.
It is booting it!

The next chapter is clear: first visible pixels of the CD32 boot animation.

I said it before, but now I mean it more than ever:
this might really be Paula32’s year.

PS - I've created a small archive in order to organise my news section.