Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: WolfToTheMoon on August 06, 2012, 12:16:46 PM
-
And not talking specifically about Amigas, but as a C64/C128 successor. Seems like an obvious plan, full backwards compatibility with their 8 bit machines, plenty fast and cheap, up to 16 MB RAM, they were granted half of the licensing cost by WDC... Yet it never happend! And I think it would make for a very powerful entry level machine in the late 80s and early 90s... certainly more impressive than C65, which to me made little sense by 91'. A GUI based OS like GEOS could have been used/licensed by C=...
Perhaps it would be dangerously close to lower spec Amigas, but I think it made sense to try and build on the foundations of 20ish millions of C64 sold and the huge software library that existed.
-
Perhaps it would be dangerously close to lower spec Amigas, but I think it made sense to try and build on the foundations of 20ish millions of C64 sold and the huge software library that existed.
A 65816 wouldn't have been compatible with alot of the existing software & commodore were happy to just keep milking the c64.
-
The C64 Super CPU (http://www.cmdweb.de/scpu.htm) accelerator has a 65816 under the hood, running at a whopping 20MHz.
Not many programs utilize it, though.
-
And not talking specifically about Amigas, but as a C64/C128 successor. Seems like an obvious plan, full backwards compatibility with their 8 bit machines, plenty fast and cheap, up to 16 MB RAM, they were granted half of the licensing cost by WDC... Yet it never happend! And I think it would make for a very powerful entry level machine in the late 80s and early 90s... certainly more impressive than C65, which to me made little sense by 91'. A GUI based OS like GEOS could have been used/licensed by C=...
Commodore couldn't have managed their way out of a hat.
Content to sit back on their money earners, they stopped innovating and, surprise surprise, a few years later they were bust.
Just this morning I was thinking that C= should have reduced the C64 to a single chip (including the VICII,SID,6502,Serial,etc) by the late 80s so that they could sell it for cheap in developing countries for a profit. It could also have been clocked higher because said chip would have been fabricated on a far more modern process than the original C64. Your idea of using the 65816 is similar. An enhanced C64 for the late 80s/early 90s, with the C64's flaws (relatively slow CPU, mud-inspired colour palette) designed out.
-
Just this morning I was thinking that C= should have reduced the C64 to a single chip (including the VICII,SID,6502,Serial,etc) by the late 80s so that they could sell it for cheap in developing countries for a profit. It could also have been clocked higher because said chip would have been fabricated on a far more modern process than the original C64.
I think the 8 bit line was actually quite well managed. They did reduce chip count and cost & because they could run the chips on their own lines they were very cheap to make. Anything more would have involved a massive investment, which wouldn't have been a good gamble.
-
Commodore couldn't have managed their way out of a hat.
Content to sit back on their money earners, they stopped innovating and, surprise surprise, a few years later they were bust.
Just this morning I was thinking that C= should have reduced the C64 to a single chip (including the VICII,SID,6502,Serial,etc) by the late 80s so that they could sell it for cheap in developing countries for a profit. It could also have been clocked higher because said chip would have been fabricated on a far more modern process than the original C64. Your idea of using the 65816 is similar. An enhanced C64 for the late 80s/early 90s, with the C64's flaws (relatively slow CPU, mud-inspired colour palette) designed out.
By the late 80s the production cost of the C-64 was about $25. It wasn't going to get much cheaper than that. Clocking it higher would have caused a ton of the software for it to not work and screwed up the video timing.
-
By the late 80s the production cost of the C-64 was about $25. It wasn't going to get much cheaper than that. Clocking it higher would have caused a ton of the software for it to not work and screwed up the video timing.
It could start up in compatibility mode, and switch to faster mode if the software desired it. It would have been a solvable issue. Anyway, all this proves is how much C= sat back on the C64 and then lost it all when the 8-bit market died overnight.
-
It could start up in compatibility mode, and switch to faster mode if the software desired it. It would have been a solvable issue. Anyway, all this proves is how much C= sat back on the C64 and then lost it all when the 8-bit market died overnight.
There was such a machine; it was called the 128.
-
There was such a machine; it was called the 128.
It was still 8 bit and pretty limited. .. I was thinking more of a next generation entry machine...a 16 bit CPU, expandable RAM, 3.5" floppy, GUI based OS...
-
It was still 8 bit and pretty limited. .. I was thinking more of a next generation entry machine...a 16 bit CPU, expandable RAM, 3.5" floppy, GUI based OS...
Sounds like an a500
-
Sounds like an a500
yes it does, but it would be cheaper and have backwards compatibility with C64 software. Plus, a 20 MHz 65816 is faster than a 7 MHz 68000 so it could a very interesting package. Timeframe around the original C65, 90-91', A500 would be replaced with A1200 in a year, A600 would not be needed.
-
Plus, a 20 MHz 65816 is faster than a 7 MHz 68000 so it could a very interesting package..
I'm not sure "faster" == "faster" in this case...
desiv
-
I'm not sure "faster" == "faster" in this case...
desiv
IIRC, the 65816 is missing some on-chip registers compared to the 6500-series, and pales in comparison to the 68000's 16 registers. This will drastically affect performance. (If you really wanna learn some gritty stuff, check out a MIPS emulator. It's fun!)
But, yeah, the 128 was still 8-bit but backwards compatible with a compatibility mode for (the dying) CP/M. If Commodore had released the C65 (comparable to the Apple IIgs in terms of "next generation") I think that would have been the end-all-be-all.
-
I'm not sure "faster" == "faster" in this case...
desiv
Well, Wozniak did say Apple demanded that the 65816 in the II GS was clocked at only 2.8 MHz to make sure it didn't threaten the Mac performance and sales. In his own words, a 7 MHz 65816 is about as fast as a 14 MHz 68000 - but I'll stipulate that a 68000 is the more advanced and more elegant solution better suited to higher level languages.
-
LOL sounds like I haven't been the only one to have wet dreams along the lines of "If I were a Commodore CEO, the Amiga would be one of the major platforms in business as well as multimedia by now" (^__~ )
-
yes it does, but it would be cheaper and have backwards compatibility with C64 software. Plus, a 20 MHz 65816 is faster than a 7 MHz 68000 so it could a very interesting package. Timeframe around the original C65, 90-91', A500 would be replaced with A1200 in a year, A600 would not be needed.
It wouldn't have had meaningful backwards compatibility, just like the C65.
In 1989 they should have had an 020 machine with chunky graphics.
I'm not sure that the 20mhz 65816 used by CMD was available that early either.
-
Well, Wozniak did say Apple demanded that the 65816 in the II GS was clocked at only 2.8 MHz to make sure it didn't threaten the Mac performance and sales..
Although, here he listed purely technical reasons:
http://www.1000bit.it/support/articoli/apple/a2gs/interview_woz.asp
Our early ideas for the computer had it running at 8 MHz. Soon we found we had to back off to about 5.5 MHz, and then to 2 MHz for that version of the processor. In the end, the product came out at about 3 MHz, which is a good compromise.
And it's hard to imagine that the 65815 is twice as fast as the 68000, although I suppose that all depends on what you're doing with it...
With the Mac, it was doing everything, so that 7Mhz 68k did feel really slow....
desiv
-
It wouldn't have had meaningful backwards compatibility, just like the C65.
In 1989 they should have had an 020 machine with chunky graphics.
I'm not sure that the 20mhz 65816 used by CMD was available that early either.
Around 90', max from WDC was 12MHz... Doesn't really matter, even a 7 MHz 65816 will be more than fast enough.
Compatibility can be improved thru emulation.
-
Although, here he listed purely technical reasons:
http://www.1000bit.it/support/articoli/apple/a2gs/interview_woz.asp
And it's hard to imagine that the 65815 is twice as fast as the 68000, although I suppose that all depends on what you're doing with it...
With the Mac, it was doing everything, so that 7Mhz 68k did feel really slow....
desiv
65816 is much more cycle efficient than 68K... as long as your using 16 bit data and relatively simple instruction, it can be even more than 2 times the speed of an 68000... however, 68000 is much faster on 32 bit ops and has more registers and more instructions... it's a better chip, no doubt, but a well optimized code for the 65816 will FLY!!!
-
Around 90', max from WDC was 12MHz... Doesn't really matter, even a 7 MHz 65816 will be more than fast enough.
Compatibility can be improved thru emulation.
You wouldn't be able to do it, if you emulated the 6502 your timing would be all wrong & you don't have enough processing power to emulate enough in software to relax the timing issues.
By 1990 the amiga was underpowered, spending money on developing another underpowered computer doesn't make sense.
-
You wouldn't be able to do it, if you emulated the 6502 your timing would be all wrong & you don't have enough processing power to emulate enough in software to relax the timing issues.
By 1990 the amiga was underpowered, spending money on developing another underpowered computer doesn't make sense.
There's no need to emulate 6502, there's a compatibility mode in 65816.
You're missing the point... yes, it wouldn't be the fastest machine around, but that's not what it's for. It would be dirt cheap and plenty fast. A early 90's C64...
Another bonus would be for WDC that would allow them to maybe introduce 65832 and 65032 chips eventually.
-
There's no need to emulate 6502, there's a compatibility mode in 65816.
It's not compatible enough. Both for timing and the instruction set.
To make it faster than an a500 would require faster memory than an a500. It wouldn't have been cost effective enough.
To achieve the price you wanted would require you to aim around the same speed as the c65/snes/iigs.
-
You're missing the point... yes, it wouldn't be the fastest machine around, but that's not what it's for. It would be dirt cheap and plenty fast. A early 90's C64...
You are missing the point. This is not just about speed - it's about price, target audience and percieved value for money.
The C64 was still selling like crazy in 1990, while the A500 had dropped below the 1000 DM/350 UKP mark by then. Between these two, there was simply no space for another machine. The C128 had been discontinued for that very reason, despite selling better than the A500.
Introducing another machine between these two would have just hurt sales of the existing two options. It would have made an A500 look to expensive or the C64 to slow. Not to mention that a third incompatible platform would have been extremely stupid, especially as late as 1990.
As somebody already said: the C64, the A500 and (to a lesser extent) the C128 were handled properly. The real problems were the lack of focus, the lack of advertising and the number of crappy computers (+4/C16/C116, A500+, A600) Commodore released, apparently to stop their best sellers from becoming too successful.
-
It's not compatible enough. Both for timing and the instruction set.
I used to have a SuperCPU, and it was pretty compatible, AFAIR. The official 6502 instruction set was fully supported, it just didn't handle illegal opcodes. But that was more of a problem when watching demos, most games and especially applications worked fine.
-
The reason is probably similar to why the C65 was cancelled before consumer production. They would have competed too closely with the 500, which the C65 was already positioned to do, and the IIGS did with the lower end Mac.
I think everyone knew these machines were dead ends so having them cannibalize sales of the 16/32 bit machines would have been shortsighted and really wasted development time.
Think, if Commodore didnt 'waste' time, money and talent on the C65 and instead put that effort into Amiga how much better the 500/2000 might have been.
I think Commodore also probably felt that with 64 Emulator 2 for the Amiga, 8bit Commodore folks had an upgrade path in the 500.
Why they chose the chip they chose for the C65 development instead of that one is probably the usual, Commodore would rather use their own chip instead of someone elses.
-
65816 is much more cycle efficient than 68K... as long as your using 16 bit data and relatively simple instruction, it can be even more than 2 times the speed of an 68000... however, 68000 is much faster on 32 bit ops and has more registers and more instructions... it's a better chip, no doubt, but a well optimized code for the 65816 will FLY!!!
I would also think that the 65816 programmers would have to struggle with the same type of segmented memory architecture issues that DOS and early windows and OS/2 programmers had to deal with when trying work with more than 64K RAM.
-
Commodore couldn't have managed their way out of a hat.
Content to sit back on their money earners, they stopped innovating and, surprise surprise, a few years later they were bust.
Hattig,
I could not agree with you more that CBM could not manage their way out of a hat but I do not agree with the rest of your argument.
The Commodore 64 was their bread winner for a long time and while they did stupid things like the Commodore 16 and Plus/4, the Commodore 128 was a worthy successor to the 64 with some nice features.
I've heard others complain that the 8 bit line should have evolved but I'm not sure where you go from the 128? I too loved the 64 and I also loved the 128. I'm not trying to give you a hard time but I pose this question to you: If you try to evolve the 8bit line, don't you start running into serious problems? For example:
1. When you start adding a more powerful CPU, graphics, etc..., when do you run into the Amiga 500?
2. Cost. Again you start upgrading the CPU/Graphics and don't you run into the Amiga 500 price point?
3. Performance. Again, doesn't the Amiga 500 make more sense given the above? If Commodore could have done a hardware/software emulator like Apple did for the Apple II to Mac folks that might have been better. We all know how poorly the software only emulators ran on a stock 500.
Also, do people care? I remember people wanting to run GEOS on the Commodore 64 emulator for their Amiga 500/2000. Does that make sense? Does using Paperclip make sense on an Amiga with an emulator?
4. In a world of 16/32 bit goodness, would the market support another 8/16 bit machine even turbo-ized? I'm not sure...
Lastly, isn't the Commodore 65 the embodiment of the above issues? Where does the Commodore 65 live? It's cost was close to a 500, not fully compatible with the 64, and not as powerful as an Amiga. How do you sell that to the public in the 90's.
Hattig, I think you ask a great question.
Cheers!
-P
-
It's not compatible enough. Both for timing and the instruction set.
it was compatible enough for Apple II GS...
To make it faster than an a500 would require faster memory than an a500. It wouldn't have been cost effective enough
It doesn't need to be faster than A500, just to be a good enough upgrade on C64/C128 and allow people to use their existing C64/C128 software...
I think everyone knew these machines were dead ends so having them
cannibalize sales of the 16/32 bit machines would have been
shortsighted and really wasted development time.
I don't see why they would be dead ends... WDC had a 32 bit version of the 65xx ready and much of the Eastern Europe and the rest of developing countries had trouble affording PCs and even low end Amigas/Ataris well in to the 90s... I got my first PC, a 486/50, somewhere around 97'... and it was used... up untill then I still used my C64.
The C64 was still selling like crazy in 1990, while the A500 had dropped
below the 1000 DM/350 UKP mark by then. Between these two, there
was simply no space for another machine. The C128 had been
discontinued for that very reason, despite selling better than the A500.
I'm talking abour a successor machine to the C64/C128, not something to slot between C64 and A500.
-
they talked about the reasons for this in the book "ON THE EDGE The spectacular rise & fall of commodore" but I cannot find the page ATM, cant remember the exact details.
-
I often wonder what would have happened had Atari and Commodore merged instead of both dying off around the same time. Atari lasted a little longer than Commodore, but not a whole lot. Atari had basically announced the Falcon, then shortly after said "well screw computers, we're going back to Consoles, here's the Jaguar, behold it's awesomeness!" then afterward faded away into obscurity.
Commodore tried the same thing with the CD32, but much like the Jaguar, there weren't a whole lot of software titles that you couldn't already get for whatever 16/32 bit computer you had.
Remember what Sega did with the Genesis? They released a converter for Sega Master System games. That's exactly what Atari and Commodore should have done for their 16/32 bit machines. Just supply some 5.25" floppy drive that had some hardware in it for emulation. Would have been a killer product and allowed established software categories to be utilized on newer systems.
Of course the problem with this is that it's too 'nice' to the consumers, and management figures most people wouldn't part with the cash to get the newer versions of conversions. But honestly, would you stay with the crappy version of Double Dragon for the C64, when you could get the far better version for the Amiga?
slaapliedje
-
Remember what Sega did with the Genesis? They released a converter for Sega Master System games. That's exactly what Atari and Commodore should have done for their 16/32 bit machines. Just supply some 5.25" floppy drive that had some hardware in it for emulation. Would have been a killer product and allowed established software categories to be utilized on newer systems.
The "some hardware in it for emulation" would have had to consist of pretty much a complete C-64 and 1541.
Sega's Power Base Converter for the Genesis was just an adapter that allowed the different sized Master System cartridges to plug into a Genesis. There was no emulation involved because the Genesis contains the full hardware of the Master System.
-
One other area that C= sadly neglected was the mobile market. I would imagine that a cheap 65xx or 65xxx design would have sold very well(LCD was rumored to have had 15 000 orders on launch, before cancellation) and that the 16 and 32 bit 65 versions would be great for cheap PDAs later.
I always believed it was a big mistake that C= never persued further 6502 development as they had great talent in MOS technologies - the future Athlon K7 designer worked on C65's CPU. Browsing the web, I have found indications that Tramiel pretty much killed any serious development shorty after aquiring MOS - some former engineers claim that a 32 bit version of the 6502 was being in the works at the time. This could have had major implications for C= as it was possible that it would have been chosen by, for instance, Apple for the Mac and further evolution of the II series...
-
perhaps simply because the 65c816 has bug
I discovered with the W65C816SXB board that it was not working correctly in emulation mode
if you try this code in emulation mode
LDA #$FF
STA $40
STA $41
LDY #$02
LDA ($40), Y
either you have an error if the machine only has 64k of memory
or if the read works you get a byte from $010001 and not from $000001
there is probably a work around by taking into account the E flag
while loading the latch (if in emulation the latch is kept a $00 it should work)
-
I discovered with the W65C816SXB board that it was not working correctly in emulation mode
Interesting thread here too.
http://forum.6502.org/viewtopic.php?f=1&t=5803
It looks intentional
or the base is specified by sixteen bits and assumes the data bank as its bank, then, if an index plus the low-order sixteen bits of its base exceeds $FFFF, it will temporarily (just for the current instruction) increment the bank. The 65816 assumes that the array being accessed extends into the next bank.
-
The 65816 was (briefly) considered for the C128 design, but a lot of people in engineering weren't exactly fans of the CPU design and software support (read: assemblers) was very immature in 1984.
Before the A2000 and A500 were greenlit, Commodore Engineering was proposing a range of different projects. One of them was a C128 follow up referred to as "BMW" and "C256", IIRC. Hayne was designing the specs, and he wanted to include a 65816, despite concerns from other engineers. That machine would not have been compatible with the 64, it would not have had CP/M and I imagine it would have been barely compatible with the 128. It got killed when new Amiga machines became the only priority.
-
The 65816 was (briefly) considered for the C128 design
Bil Herd tells a different story. That he received a call from Bill Mensch trying to get him to use the 65816, which he immediately refused.
It wouldn't be commodore 64 compatible & it would be a chip they had to buy from an external company.