COMP: Re: tech snippets

From: Mark Grant (mark@unicorn.com)
Date: Thu Dec 26 1996 - 14:33:04 MST


On Wed, 25 Dec 1996, Dejan Vucinic wrote:

> someone said:
> > Taking this one step further: why not putting the entire microkernel/VM
> > into a _normal_ on-die SRAM

Does this gain you much? I was under the impression that the microkernel
only gave you very limited non-portable facilities so that you could
rapidly port most of your OS by rewriting the microkernel. I can't believe
that all your device drivers fit in 12k. How often is the message-passing
code called?

> >, having an address in address space (cache
> > has none, it occupies the same address space as the addressable core)?

Some chips let supervisor-mode software lock lines in the cache so that
they always map to one specific address, effectively allowing you to use
it as on-chip RAM. I'm fairly sure that the T9000 transputer can do this
(I read the manual a few years ago), but I'm not sure about more common
chips.

> Taking this yet another step further:
> Over the last fifteen years all major CPU makers have migrated from
> a monolithic design to microcode.

Are you sure of this? I thought the point of RISC was that with a
simplified instruction set you eliminated the microcode and could increase
the speed of your chips because the decoding overhead was reduced.

> Why in the world doesn't Motorola
> or MIPS or whoever implement both x86 and 68000/PowerPC/MIPS instruction
> set in a single chip!?

Wasn't there a rumor of a combined PowerPC/486 chip from IBM a year or two
ago? If Intel don't use microcode then you can probably forget any such
combination, because it would be a huge chip. Even if it did then the cost
of developing a Pentium clone would probably be prohibitive. I can't see
Intel supporting the PowerPC chip by giving Motorola rights to the Pentium
architecture, though they might be able to license the Cyrix or AMD
clones.

> Switch between the two with a protected
> instruction. Run DOS on an SGI machine as a native OS, for
> applications' sake.

Even with a chip that can run Intel instructions you would still need to
write a huge amount of code to emulate the rest of the PC hardware, or
license Windows and DOS so that you could rewrite them in native code.

> Bury the
> inferior Intel architecture.

I'm not sure what's so inferior about it, except perhaps for floating
point performance. As a programmer I much prefer the PowerPC architecture,
but the Intel chips are fast and cheap in comparison. The theoretical
advantages don't seem to matter so much in practice.

        Mark

|-----------------------------------------------------------------------|
|Mark Grant M.A., U.L.C. EMAIL: mark@unicorn.com |
|WWW: http://www.c2.org/~mark MAILBOT: bot@unicorn.com |
|-----------------------------------------------------------------------|



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:35:56 MST