[net.micro.68k] 68040

dillon@PAVEPAWS.BERKELEY.EDU (Matt Dillon) (06/11/86)

	128 bit internal busses are not likely for high level assembly.
It is more reasonable for low level microcode (maybe they were talking about
microcode in wherever_you_got_the_gossip_from)

					-Matt

stuart@ihlpf.UUCP (Ericson) (06/11/86)

> 
> 	128 bit internal busses are not likely for high level assembly.
> It is more reasonable for low level microcode (maybe they were talking about
> microcode in wherever_you_got_the_gossip_from)
> 
> 					-Matt

I don't know where he got the info, but I saw such a piece
of rumor in a recent "EE Times".  It said 32-bit external
and 128-bit internal.  I'm pretty suret that they said
that the registers were going to be 128 bits wide.

-- 
Stuart Ericson                       Disclaimer: My computer wrote
                                                 this - I just didn't
                                                 have the heart to
USENET: ..!ihnp4!ihlpf!stuart                    turn it off.
voice: (312) 979-4288

"This time I'm switchin' to glide..."

markp@valid.UUCP (Mark P.) (06/26/86)

> > 
> > 	128 bit internal busses are not likely for high level assembly.
> > It is more reasonable for low level microcode (maybe they were talking about
> > microcode in wherever_you_got_the_gossip_from)
> > 
> > 					-Matt
> 
> I don't know where he got the info, but I saw such a piece
> of rumor in a recent "EE Times".  It said 32-bit external
> and 128-bit internal.  I'm pretty suret that they said
> that the registers were going to be 128 bits wide.
> 
> -- 
> Stuart Ericson                       Disclaimer: My computer wrote
>                                                  this - I just didn't
>                                                  have the heart to
> USENET: ..!ihnp4!ihlpf!stuart                    turn it off.
> voice: (312) 979-4288
> 
> "This time I'm switchin' to glide..."

A more likely scenario is 16-byte cache lines (128 bits, get it?).
An unsophisticated EE Times interpretation of such a rumor would
probably get printed as "128-bit internal architecture", and this has
got to be a much more likely possibility.  128-bit registers are
just NOT useful enough to warrant the extra carry-lookahead delay,
space, etc.

			Mark Papamarcos
			Valid Logic Systems, Inc.
			{hplabs,pyramid,ihnp4,...}!pesnta!valid!markp

disclaimer: rumors are rumors, especially when filtered through trade rags,
	    and the above are only my opinions.