[mdlug] hanging machine kernel panic - on going

[Aaron Kulkis]

Ingles, Raymond wrote:
>> From: Aaron Kulkis
>> I won't bore you with the details,
>> other than to say that there are 8, 16, 32 (and now 64)-bit
>> versions of four registers A, B, C, and D (which Intel intended
>> to stand for Accumulator, Base, Code, and Data), but which,
>> other than the A registers, seem to be used rather inconsistently.
>
> Yeah, the x86 instruction set is famously non-orthogonal. Instructions
> can and do affect each other in awkward ways.

ah yes, the non-orthgonality, forgot about that one.


If you has ever programmed in VAX-11 assembly, you would probably
agree that the VAX-11 instruction set is

* Beautiful

* Elegant

* Masterful

* Sublime

* Wonderful to program in

* And totally WRONG :-(

[Primarily because it was possible for a single instruction
to produce as many as 13 page-faults, each time requiring that
the all register and memory changes be rolled back to the
contents before the instruction started executing]. A heavily
loaded VAX thrashes alot (swapping virtual memory pages in
and out) to the point where ~90% of CPU cycles are spent on
page-swapping activity and only ~10% running user code.
I cannot begin to express how painful it was to try to
do anything on a heavily loaded VAX.


It was the complete failure of the VAX to hold up well
in multi-user multi-tasking environments which led to the
development of Reduced Instruction Set Computer (RISC).

>
> Even more silly, 'under the hood' it's all translated into a
> more-or-less RISC format for actual execution. In a sense, *everything*
> on an x86 runs as a 'virtual machine'. :-/


It's such a shame that this pathetic architecture has
taken over the world :-(.  I'd rather that the dominant
architecture were something more like the 680x0 or
IBM 370 lines of CPUs (although I'm not familiar with
the Power CPU (IBM-designed microprocessor) line.

It was the complete failure of the VAX to hold up well
in multi-user multi-tasking environments which led to the
development of RISC.


>
>   Sincerely,
>
>   Ray Ingles                                    (313) 227-2317
>
>   "Any technology distinguishable from magic is insufficiently
>           advanced." - Anonymous' restatement of Clarke
>
> http://www.compuware.com/
> The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it.
>