Originally posted by MacMarshall
I thought that what killed Moto in the CPU market was when Apple decided to move from CISC to RISC in 1993, which killed Moto's 68050 processor. Intel since proved that CISC is scalable after all.
But maybe you have more details than that.
Well it goes much further back than that. Back to a different time and age (picture your screen going hazy and wavy around the edges).
Back in the early 80's Mot ruled the high end cpu landscape. At the low end you had the Z80 and 6502. In the "middle" you had x86 and 68k, and on the high end you had the 68k. Now by high end, I'm referring to computers. Note that back them, pretty much the ONLY computers running x86 were pc clones. Most of the workstation manufacturers were using 68k (Sun, HP, DEC, SGI, etc). Then this RISC thing came along, and Sun (who was even back then the big boy in workstations) went to Mot begging for them to come up with a RISC chip. Mot declined/drug their heels. Sun gave up in frustration and went forth with their own design (SPARC). By then HP had PA-RISC, DEC had MIPS then Alpha, SGI had MIPS, and IBM had ROMP then RIOS/POWER. Motorolas hold on the workstation market evaporated. Way too late Mot came up with the 88000, which was a nice chip, but by then all the big boys had theirs and there weren't any more major manuf. left to use the chip. They did manage to get Bull and Data General to sign on. DG came out with the Aviion line. It was rumoured that NeXT had a dual proccie 88k motherboard to supplant the 68k, but then they dropped hardware and that was the end of that. Sun also used the 88k in some deskside graphics subsystems. (As a side note AMD suffered a similar problem (late to market) with their RISC effort, the 29000. It did see some success however as an embedded controller, used esp. in laser printers).
Fast forward a little and Apple realizes that they are being left in the dust performance wise and needs to do something "radical". In steps IBM and Mot. 88k lives on a bit in that part of the PowerPC uses tech from the 88k (bus and glue logic I hear, but I don't know the details). BTW, I believe that Mot was working on the 68060 at the time, not an '050 (now that's interesting eh, we're discussing the same skipping over of '5' with G4->G?). Some '060s shipped in products (I know some Amiga accelerator boards shipped with them). Performance wise, the '060 would probably hold it's own if not beat the early 601, but by then it was too little too late. Even back then Mot was having issues ramping up the clock speeds in the 68k series (see, some things NEVER change).
So had Mot played their cards right, they themselves (all alone without IBM's help) would be the #1 selling RISC processor, powering Sun workstations/servers and Apple, plus probably several other workstation manufacturers machines. Of course this pales compared to where Mot would be now (and conversly Intel) had IBM picked the 68K over the 8088 for the PC, but thats another story
As for Intel and CISC. Yes and no. The current line of x86 processors are really RISC processors at their core (literally and figuratively). So they proved that you could design a chip that executes CISC instructions quickly, but that it takes basically a RISC "chip" to do it.
And finally, Pentium being 586. Well that was true way back when, but had Intel stuck with the numerical model names (which they changed for marketing/legal reasons), you can bet that todays P4 would be way past 586. 986 if you take (Pentium == 5) + 4.