*updating* 8088's

jpero at sympatico.ca jpero at sympatico.ca
Tue Nov 20 14:44:33 CST 2007


> On Nov 20, 2007 11:34 AM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
> > > I'd like to see an 80286 to 80486 upgrade card. It would need to somehow
> >
> > I have one. It's currently running in the machine I am typing this on.
> >
> > It's a little module with a TI 486-compatile processor on one side and,
> > IIRC, the floating poiut coprocessor for it on the other (or at least
> > that's the only explanation I can find for 2 large PQFPs). A couple of
> > PLDs and a PLCC 'plug' to go into the 80286 socket.
> 
> Is it a true 486? or a 486SLC (which is really a Cyrix 386SX with an
> internal cache and a couple of the 486 instructions included)?  IBM
> and TI were second sources for the Cyrix processor.  But it really
> wasn't instruction compatible with the 486, and most instruction
> timings were more 386 like than 486 like.  It didn't work like a 486
> especially where cache control was involved, even though it had 486 in
> the name.  But the cache, and better integer multiply circuitry, made
> it significantly faster than a 386SX of the same clock speed.  That
> is, assuming you could run the program to enable the cache, or had
> BIOS support for the cache.
> 
> Cyrix also made a 486SRX2 which was clock doubled, as well as the
> 32bit versions, the 486DLC and 486DRX2 which would work in a 386DX
> socket.
> 
> Anyway, the SLX or SRX2 (which have 386SX pinouts) would be much
> easier to adapt to an 80286 socket than a true 486 would be.
> 
> Eric

Eric and others:

Ever you had IBM sourced 486 SLC/SLC2/SLC3?  Far better than Cyrix/TI 
486SLC wannabes.  Cyrix/TI utilized 1K cache.  IBM had 8K and 16K and 
much better optimizations.

Cheers, Wizard 



More information about the cctalk mailing list