Imagine, how things could have been...

[bloodline]

It's no secret that From a technical stand point IBM's choice of the 8086 for their "brand new" Personal Computer back in the early 80's was a shock to almost everyone.

All other manuafactures were looking at the 68K as the CPU of choice, simply because it was a better chip. IBM, contract with intel to supply 8086s very cheaply, apparently had a heavy bearing on the CPU choice.

But, I wonder what would have happened if IBM had chosen the 68k....  :-o


Just some thoughts to get the ball rolling:

Motorola would not have had the ruthlessness of Intel, and desire to improved the Chip. Right now we would all be using 100Mhz 68040s...
IBM would not have gained such a strangle hold on the market as there were plenty of other, better machines avaiable which, with a little poking, were quite compatible... :-D

[bloodline]

DaveP wrote:

Motorola would not have had the ruthlessness of Intel, and desire to improved the Chip. Right now we would all be using 100Mhz 68040s...

Highly unlikely.

you're right... we'd be on 20Mhz 68000s....  :-D

[bloodline]

The 680x0 line really didn't have much in the way of clones. If there had been, those manufacturers using 68K would have sourced the best 68K price/performance ratio and maybe we would have seen 68K chips with the same lofty clockspeeds as todays' x86

It appears Motorola were not fussy about who they gave licences to... so I don't expect there would have been the same sort of clone wars the Intel has had to face.

I mean, Intel has had always had to face other manufactures cloning their chips (from teh Z80 onwards), putting them in direct compretiton... Motorola on the other hand seen to have been happy to licence the technology, thus there would be more regulations on the technology, and less "push", I certainly don't think the PPC would have ssen the light of day... and the 8086 would have die without trace. The dominant chip now would be the Alpha...

[bloodline]

Matt_H wrote:

Imagine, how things could have been...

No. Let's not. It makes me want to cry.  :cry:

But seriously, that deal probably would have made the 68k the mainstream chip. Therefore, it's logical to assume that the product line would have been expanded, and we'd probably have 68090s by now.

Amiga Format ran a news article once about theoretical speeds of an 060. They went into the 200MHz+ zone with refined manufacturing. There's a lot of technological potential left in 68k, just not financial.

Technologically the 68K and the 286 are from the era (with in reason!?!?), so one can safely assume that if the x86 can be pushed to the Athlon and the P4... there is no technical reason why the 68k can't be pushed in the same way.

I'm somewhat surpprised there has been less interest in 68K development.. History has shown that CISC/RISC hybrids (060,Pentium,Athlon) are generally more scalable, versitile  and compatible than their RISC cousins (PPC,Itanic...)

Still it's a shame about the Alpha  :-x

[bloodline]

iamaboringperson wrote:
just having a 68000 wouldnt necessarily make the machine as a whole equal to the mac/lisa/amiga/atariST/etc...

imagine if they did that - but had no custom chips - still used an MS-DOS type thing, and still had all the interupts and crap(quite possible), its still quite likely that the wouldnt have an automaticly configuring machine

the 640K limit would be gone! that would be good

it would of been interesting to see

but the amiga & mac(and many others) still had advantages over the ibm-pc other than the CPU
so perhaps in terms of sales and support and marketing and popularity - i dont really think it would of changed that much

Indeed I agree, PC's may well have still dominated, but the CPU market would be totally different, and probably we wouldn't have such great CPU's now...(Nescesity being the Monther of invention)

One thought though... If the IBM PC had used the 68k, then the Amiga/Mac/ST, could all have run "PC" software in a Virtual machine "Dos Box"... that would have shaken up the market a great deal, I think.

[bloodline]

mdma wrote:
If it was possible to cool an 060 with something like liquid-nitrogen, what would be the theoretical maximum clock speed achieveable?

About 80Mhz :-D

-EDIT- You'd probably need liquid Helium (Chemistry/Physics joke) to cool it down enough.

I think the 060 was at it's limit at 80Mhz though, as you then get cross talk, and delays between the gates which cause the whole thing to lock up. I would need longer pipelines, out of order scheduling, more advanced branch prediction, register renaming and a massive register set coupled with a nice big L1 cache (Basily all the tricks squeezed into the Athlon) to get it to go any faster...

[bloodline]

patrik wrote:
Talking about the limits in terms of the 68060's clockspeed.. has anyone tried raising the Vcc supplied to the 68060 to reach higher clockspeeds?

Normally when overclocking, if you are aiming at high clockspeeds  you have to raise the Vcc supplied to the processor. In the 68060 User Manual it says that the absolute maximum rating of the 68060's Vcc is 4.0V. It would be very interesting to see what kind of results a Vcc of maybe 3.8V would give... asssuming that the memory etc would cope with the higher clockspeed.


/Patrik

I think you hit the nail on the head, the current 060 solutions will probably max out due to the Ram and supporting BUS logic before the 060 reaches it's upper limit.

[bloodline]

mdma wrote:

iamaboringperson wrote:
ive always wanted to set up a pentium 4 machine and run win3.11 on it

I installed DOS6.22 and Windows 3.0 on a P100 with 32MB EDO RAM about 5 or 6 years ago for a laugh, it booted in about 3 seconds and was fast as f*ck!  It'd fly on a P4, a bit like AROS does really! ;-)

Yeah, the AROS Gfx demos give you a good idea of how much power is reving under AROS... The Gfx drivers are a total bottle neck at the moment, they have no acceleration what-so-ever. When the PCI drivers improve I think the gfx drivers will mature and you might even poop your pants at the speed :-D

The IDE drivers are really crappy too, PIO-0.. Can't wait to get DMA on those  :-D

[bloodline]

mdma wrote:

bloodline wrote:

mdma wrote:

iamaboringperson wrote:
ive always wanted to set up a pentium 4 machine and run win3.11 on it

I installed DOS6.22 and Windows 3.0 on a P100 with 32MB EDO RAM about 5 or 6 years ago for a laugh, it booted in about 3 seconds and was fast as f*ck!  It'd fly on a P4, a bit like AROS does really! ;-)

Yeah, the AROS Gfx demos give you a good idea of how much power is reving under AROS... The Gfx drivers are a total bottle neck at the moment, they have no acceleration what-so-ever. When the PCI drivers improve I think the gfx drivers will mature and you might even poop your pants at the speed :-D

The IDE drivers are really crappy too, PIO-0.. Can't wait to get DMA on those  :-D

Still faster than Gayle though! :-D

Yeah, but the Royal Mail is faster than Gayle :-)