A short but important note from the Buffee project. Things is moving forward! Here is the latest:
Just a small update; we’ve been working hard at getting the core of PJIT ready to boot up before receiving our first batch of boards. Getting the AM3358 up and running without the benefit of JTAG has been … interesting. To date we have: – working clock reconfiguration – set up of all the caches and mmu – a timer by which we can measure performance – our first runs of PJIT with the BogoMIPS test Initially, our very first run was pretty abysmal, but our BogoMIPS score is now between 400 and 666 using the pedagogical BogoMIPS for the 68030 (which ought to take precisely 4 cycles per loop to complete — this isn’t a compiler or 68K optimization exercise).
I am looking more and more forward to this. With the PJIT testing started this project is on the move which is important.
For comparison, the CPU natively gets 1000 BogoMIPS (or 990 in Linux, because Linux sucks), which means our emulator is between 40% and 67% of native performance! Right on target!
Progress! Kudos to the work being done here for the Amiga and the rest of the retro machines and consoles that this product is aiming for.
I was told it couldn’t be done, and for a while, I was worried they might be right. But they’re not and the PJIT model totally works.
At this point nothing shocks me anymore. I remember standing on the floor at The Gathering in 1996 when The Black Lotus presented Tint demo. My heart pumped and my hair raised. It was too good to be true? Then we have all the creative hardware milestones that been happened on the Amiga by 3rd part companies and from persons dedicating their life to this amazing computer.
Once you hit a barrier in the Amiga community,… it seems like to actually reveal and shock is not so unusual. Buffee seems to be on right track and with these words it is a fact. Buffee is happening. The speed-kick for all OCS Amigas and more!
It does depend on aggressive inlining and smart selection of which eight of the sixteen 68K registers we keep in the CPU and which end up in SRAM. But I’ll be making that a compile-time option so beta testers can play with any combination we like.
I see that the project have seen many interests in being beta-testers. They are all vital for the success of this project. Testing and reporting is vital to get a final product that will make tons of people happy.
Our next step will be implementing a little debug shell where the user can peek and poke around memory as well as getting the GPMC up and running with what we believe are the correct timings. Stay tuned.
Take your time. No rush. This is from the Discord channel. But you can also read it up on the Buffee homepage. I will be here covering the Buffee project as long it takes. A nice CPU + MEM accelerator for the 68000 socket. Your Amiga 500 will never be the same once this project is done. What can AmigaOS 1.3 achieve with a 3.2GHz CPU?
What sort of Amiga upgrade is important to You?