🚀 go-pugleaf

In article <BxHFvo.K0B@ux1.cso.uiuc.edu> news@ux1.cso.uiuc.edu (News) writes:
[...]
   Q: "Will NeXTSTEP 486 run on the Cyrix 486SLC?"
   A: "No. The Cyrix chip is not a true 486."

       That shows there is definitely a big difference between a Cyrix 486SLC
   and a Intel 486SX/DX.  Otherwise, NeXT would advertise support for the
   Cyrix chip.
Not necessarly big, but something essential...
       The info sheet does say that NeXTSTEP 486 will support ISA/EISA 486SX,
   486DX, 486DX2, and future members of the Intel x86 family.  The IBM 486SLC
   is not addressed, but that's probably because NeXTSTEP 486 won't support
   MicroChannel.
I've heard of a 386SLC, but a 486SLC? What's that?

   It clarifies by saying that NeXTSTEP 486 will run on any
   TRUE 486, so AMD better do a good job or they'll be left out too.

   Jeff Barcalow

Now just assuming that Intel isn't bribing or secretly threatening the
software industry (not impossible I'm beginning to think--where is
anti trust hiding out, anyway?), there must be technical reasons why
NeXT-Step 486 won't support the 486 (if it really doesn't support the
386. After all, while NeXT Step isn't SUPPORTED on the 386
(performance reasons :-), it might still run on it).

The only technical reasons I could think of are:
- the 386 is too slow
- NeXT Step uses PostScript, PostScript means floating point (but a
486SX doesn't have it either...hmm)

- NeXT Step 486 relies on page faults at ring 0
NeXT Step is supposed to be based on MACH, MACH runs on a 386, so
unless they've changed a lot there, I don't see why they should
require that. I could see that they want to save ring transition
overhead in a micro kernel system (so they have system processes run
at ring 0 but with some protection like read-only pages and a pageable
kernel), but is it worth excluding the whole 386 market (laptops)?

So are they using 486 just to say that they are better than the
average 386 Unix or what?
---
Thomas M. Hoberg   | Internet: tmh@first.gmd.de
1000 Berlin 41     |           tmh@cs.tu-berlin.de
Wielandstr. 4      |
Germany            | BITNET:   tmh@tub.bitnet 
+49-30-851-50-21   |