MEGA65 FORUM

Proper I/O in native mode, any chance for GPU?

append delete Ratteler

One thing I always hated about the old 8 bit Commodores was the mirrored I/O of all the chips.
The fact is the only reason that existed was to save some cost on millions of units, and in 1982, no one could imagine using up 64k.

Since the C=65 will never exist as more than a prototype, and the Mega65 doesn't actually exist yet, and we won't save anything by NOT properly decoding all the I/O, we can free up a bunch of I/O space in native mode.

Also, I assume the GS4510 is 65816 instruction compatible as the C=65 would have been 65816 based. (This actually confuses me since in native mode it would have been the first of the 16-bit Commodore line in the C64's lineage.)

If we can free up enough I/O through removal of mirroring, maybe we can add this.

http://miaowgpu.org/

An actual Open Source GPU for 3D graphics and FPU Calculations.

Reply RSS

Replies

append delete #1. MIRKOSOFT

Where you found that 65CE02 (4510) is 65816 opcode compatible?
This is really big mistake and not true.
I'm owner of SCPU128 and I'm programming on 65K platform, really these CPUs (here it needs to say that 4510 is MPU, 65CE02 is CPU) are different and incompatible except X65 emulated mode. Native 16-bit mode of 65816 is different and at correct configuration can run X65 code. Of course true is that 65816 and 65CE02 both supports not illegal opcodes.
Still in emulation mode, in 65816 native mode and full usable instruction set of 65CE02 are incompatible. Mega 65 has in progress support of illegal opcodes of 6502 in C64 mode, this is not real in case of 65CE02 even 65816...
If can be Mega 65 based on 65816 it can be different computer compatible with my SCPU128....

Miro

:: @MIRKOSOFT added on 04 Apr ’16 · 23:06

I forgot - I'm 128er and C128 allows by MMU and its bankswitching use whole RAM also RAM occupied normally by I/O.

append delete #2. Ratteler

I just assumed since they were trying to replicate a C65, they would use an FPGA core of the processor that C65 was based on. Apparently I was wrong.
I'm slowly making my way through the github.

Instead, it seems like they are crippling the advanced abilities of the C65 for C64 compatibility. Something that many in Commodore's own engineering were dead set against at the time they were working on the REAL C65.

The argument was that the 128 was not as successful as it could have been because no one wanted to code for 128 mode when 64 mode was highly compatible and they could reach the whole market including the 128 by staying in 64 mode.

There was talk of severely crippling 64 mode to force developers into C65 native mode.

This is why I'm glad this project is open source. Progress on stuff like the VIC IV can be recycled into a truer C65 style machine as a fork in much the same way people are still making boards to use real SID chips instead of emulating them.

I'm beginning to understand that they truly are making C65 inspired 8 bit machine instead of the true 16 bit Commodore that those who believed in the project had hoped it would be.

We should nickname the Mega65 the Mehdi65. :-p

append delete #3. MIRKOSOFT

In case 128 you know not about software for it?
When we look at 128 - it was not game machine, so if you're talking about games, it's true - software for 64.
Also 64 mode was the biggest disadvantage of 128 - VIC2E for C64 compatibility did slow clocking of Z80, possible use only 1 MHz of 8502 on VIC screen, at least approx. 1.3MHz is possible to get, Z80 instead 4MHz downclocked to 2MHz...
But advantages: 2 desktops, 2CPUs, fast serial...
Currently is in discussion on 128 forum VDC resolution 1024x256...
In case C65 we cannot mark 64 mode so big disadvantage 'cause no software, no real experience of using C65 native mode could be created for unreleasing computer...
Miro

append delete #4. Ratteler

I'm just talking about history. I was a beta tester for CMD and even worked at their booth at the 1990 World of Commodore Amiga show in NYC. The show was rebranded Amiga only after CMD bought a both.
I was running animation at the booth of a 3D CMD HD off the RAMlink, and had to repeatedly explain... "It's coming from a C=64." when people who passed by were critical.

CMD was contemplating what peripherals they could sell for the C65. They were in talks with Commodore about making the Atari style parallel interface they used between the RAMlink and the CMD HD standard on the C65 for better drive access. So at the time I had a tiny bit of third hand inside knowledge on the project.

So the comments weren't mine. They were echos of the concerns of people inside Commodore at the time the C=65 was in development.

We were ALL real excited about the prospect of a 65816 based Commodore. Me in particular. I was actually one of the original forces behind the concept of a ZIPChip style accelerator for the C64 based on a the 65C02 and actually kind of inspired the SuperCPU, through I never owned one.

https://archive.org/stream/commodore-world-20/Commodore_World_Issue_20#page/n16

I am the "Joe Tom Collins" mentioned in the article. FYI.

That's why the SuperCPU went with the 65816. CMD had already done a lot of research on pushing the chips capabilities, in prep for the C=65 and was considering a CMD 65, for want of a better name, as Commodore was already past it's deathbed vigil by this point. So they recycled that research and made it an accelerator board for the existing 8 bits, overcoming some really absurdly complex obstacles that Jens Schonfeld of Individual Computer won't even TRY to address with his Turbo Chameleon.
I think he just didn't like the 128. He actually says plugging it into a C=128 will VOID YOUR WARRANTY!
If it wasn't for that, I would own one of those.

We were actually dreaming of a future with the WDC65C832 which never materialized as more than a data sheet.

I don't know what version of the C=65 the Mega65 guys are trying to recreate, but it's sounding less and less like the machine WE, and Commodore's engineering team were dreaming of back in the early 90's.
Which is perfectly OK. The fact that they Open Sourced the project is awesome, and FPGA's are now cheap enough and complex enough to allow any one of us to design the computer of our dreams.

append delete #5. MIRKOSOFT

Ok, really surprised... You're man from CMD...
I'm not to going declare which CMD devices have/had, but only one I need to talk about - SuperCPU.
I own model for C128 and know its advantages/disadvantages... Really have so many Qs which are not for this forum, so, can I give you my email address to ask you anything of them?
I don't want to forget this forum, but here's no place for C128, SuperCPU and devices not related in M65.
Miro

append delete #6. gardners

Thanks for the missing pieces of history there Joe -- very interesting to hear.

So, all the C65 prototypes that I am aware of have a modified 65CE02 (4502), not a 65816 as the CPU. I had never seen anything about using a 16-bit 6502 variant, although it would have made a lot of sense of course.

That said, the 4502 does include some 16-bit instructions, so I am not really sure what was going on.

The main crippling I see in the C65, is that the bitplanes were royally crippled, presumably to prevent competition with the Amiga. It seems that some of the C= engineers tried to overcome this in the 2nd version of the DMAgic chip, which includes some nice bitplane scrolling acceleration.

However, in terms of where we stand today, the 45GS10 in the MEGA65 runs at 48MHz already, and while not as convenient to program as the 65816 for somethings, it is almost certainly at least as fast as a 65816 for most work loads, if only because we have the higher clock speed on our side.

As for C128 compatibility for the Chameleon, I must say that I sympathise with Jens on this point: The C128 makes life MUCH harder for this kind of cartridge. Of course this is also a compliment from me to the CMD guys who managed to make the SuperCPU work on the C128. I'd have simply given up, or instead designed a whole new motherboard where you just pull all the C= chips out of your C128 motherboard, and replace the whole thing ;)

Paul.

append delete #7. Ratteler

I'd have simply given up, or instead designed a whole new motherboard where you just pull all the C= chips out of your C128 motherboard, and replace the whole thing ;)

Paul.

That was actually my idea to the CMD crew before they decided to go the SuperCPU route. :-D

I jumped ship for the Amiga before the SuperCPU's were available, but I've heard there was talk of a CMD "GUS" Computer that never materialized.

append delete #8. gardners

Interesting...

Reply

(Leave this as-is, it’s a trap!)

There is no need to “register”, just enter the same name + password of your choice every time.

Pro tip: Use markup to add links, quotes and more.

Your friendly neighbourhood moderators: Deft, gardners, MARCOM