68K's had FP coprocessors too
That was Motorola's products.
There were no reasons for Amiga not to go in the 3d accelerator field, and chunky modes were in AAA
CSG/MOS would require some talent in building a fast RISC processors. During late 80s, SGI uses Intel 860 RISC/3D hybrid as its 3D accelerator.
Today’s 3D accelerators are FP processor arrays. Both ATI and NVIDIA have engineers from SGI.
Intel processors are fast now, but they have the burden of a 30 years old 8086 to carry as compatibility
Only a minor burden since X86 ISA occurs at the front-end of the CPU i.e. hardware/micro-code emulation/translation from variable length instruction (CISC) to fix-length RISCy ISA.
This burden didn’t stop AMD and Intel adding SIMD, Out-Of-Order, super-pipelining (includes FP), Fused FMUL/FADD(C2D), RISC-core, quad-instruction issue(C2D), speculative instruction(C2D) and data prediction and any other DEC Alpha EV6 features.
The actual CPU implementation is a customised RISC core.
The variable length instruction (CISC) has an added benefit of instruction compression.
A modern X86 processor operates in modes. AMD64 (aka X64, Intel64, EMT64) killed 'Real Mode' 8086 compatibility.
They are also faster because they are produced in big volume
Unit sale numbers doesn’t inherently equals faster CPU cores e.g. ARM and MIPS.
What’s important are the people who designs these CPUs.
I would love to see what would happen if a better architeture like ARM or PPC had dominated the PC industry. We would be way ahead of our time.
Won’t change much since the people who designing today’s X86 would be still building MIPS or Alpha or PA-RISC.
Industry players would be the same minus the X86 ISA front-end.
For example;
"The cache design of the AMD Athlon is very similar to that of the Digital Alpha 21264(EV6). The repeated use of Digital Equipment Corporation(DEC) techniques which have been licensed by AMD can be explained by the fact that the development of the Athlon was led by Dirk Meyer who was head of development of the DEC 21264 at theDigital Equipment Corporation labs."
Since AMD’s K7 Athlon shares the [d]similar[/d] same infrastructure as with DEC’s Alpha EV6, just replace Athlon core with Alpha core and run Windows NT.
The last known Windows NT build for DEC’s Alpha was Windows 2000 (NT5.0). This version was used internally by Microsoft to build Windows NT 64bit editions.
In late 90's 1Ghz race, both Intel's Pentium III and AMD's K7 Athlon has a similar high clocking nature (when compared to PowerPC group) as DEC's Alpha EV6. One of the major reasons was the break-up of DEC and subsequent spread of talent to Intel and AMD.
In the alternative timeline, PowerPC gets beaten up again in early 90's 1Ghz race, via DEC’s Alpha (or any neo-DEC teams).
IF CBM was alive, we might be using HP's PA-RISC and ultimately Intel’s Itanium (includes X86 compatibility).
Refer to
http://en.wikipedia.org/wiki/3DO_Interactive_Multiplayerhttp://www.system16.com/hardware.php?id=575Lets put
1. 3DO vs ATI (Mach32) in 1993 context....
2. 3DO's M2 vs 3DFX(Voodoo2)/NVIDIA(TNT)/ATI(Rage 128 GL) in 1998 context.
