[rescue] Perverse Question

[Dave McGuire]

On Sunday, June 15, 2003, at 07:25 PM, Carl R. Friend wrote:
>>>    I know for a fact that some of the machines under my charge
>>> at work run frighteningly hot (better than a 40 degree rise from
>>> air intake to exhaust) and I'm sure won't have terribly long
>>> lives.  This does not make me happy.
>>
>>    Wow...40 degrees?!
>
>    It may not be *that* much, but it is rather alarming to walk behind
> the rack where two of the machines are located and have a very "hot"
> spot in an otherwise air-conditioned room.  I'm guesstimating that
> the exhaust temperature is around 115 (F) or so.

   I'd say 40 degrees is quite a big intake/exhaust difference for a PC, 
man.  But that's just me. :)

>    I remember reading a trade rag of some sort touting "Asynchronous
> Logic - The Wave Of The Future" a little while ago.  How soon we,
> as a species, forget....

   This is exactly why I spend so much time studying the history of this 
science.  It's fascinating to read about what was tried in the 1950s 
and 1960s. :-)

>    I'll concede the latter point, although the 11/xxx machines did
> (damned near typed "dud" there) have "compatibility mode" built in.

   That they did.  PDP-11 compatibility mode is an "official" part of 
the VAX architecture, but according to the VAX Architecture Reference 
Manual its implementation is listed as optional.  This may be a matter 
of opinion, but I'd say the ability of a processor to kick itself into 
a mode in which it can execute the instructions of a different 
processor doesn't tie into the lineage argument much.  I could be 
easily swayed from that opinion, though.

>>    Even at the hardware/implementation level...a complete 8080-based 
>> CPU
>> consists of three chips: the 8080 CPU, the 8224 clock generator, and
>> the 8228 system controller and bus driver.  A complete 8086-based
>> system consists of an 8086 CPU, an 8284 clock generator, and an 8288
>> system controller and bus driver (the latter of which is admittedly
>> optional for the 8086).  While they're not binary-compatible, 8080 vs.
>> 8086 assembly code is nearly directly upward compatible.
>
>    This is a common thing.  Usually no microprocessor stands alone;
> they need a certain amount of "glue" to bind them into a system.

   I wasn't saying "this is a chipset and that's a chipset, so they're 
the same"...Look at the datasheets for the six chips discussed.  
They're nearly identical in all respects, when examined 
signal-by-signal.

   Architecture aside (and I realize we were discussing architecture and 
this is a digression) the implementation of the 8086 chipset is 
basically a widened 8080 chipset.

>   A
> few vaguely similar part numbers and identical assembler op-codes do
> not make for a properly proveable "parentage".  (See above for the
> Interdata bit.)

   Your Interdata point is a good one...but the similarities here run 
much, much deeper.  As a start, look at the register sets:

   8080		8086
   ----      ----
    A		AH/AL (AX)  (same, just 16 bits wide instead of 8)
    B		BH/BL (BX)  (ditto)
    C		CH/CL (CX)  (ditto)
    D		DH/DL (CX)  (ditto)
    E					
    SP		SP
    PC		PC
			CS			(four segment registers...to extend the
			DS			 16-bit 8080ish address space to 20 bits
			SS			 using a crude and non-programmer-friendly
			ES			 segmentation scheme)
			BP			(wow, some NEW architectural features here)
			SI
			DI

>    All that bile aside, however, I'm sure that the 8080/8085 strongly
> influenced the design of the 8086.  But to call the 8086 a direct
> descendent of the 8080 seems, to me, to be a bit of a stretch.  To
> me a "direct descendent" is more akin to the PDP-1, PDP-9, and PDP-15
> lineage of machines, or the entire Nova/Eclipse line.

   It was the mid-1970s, man...everything influenced everything else in 
the brand-new microprocessor world because nobody had any clue about 
what they were doing.  The 8086 was clearly a new processor...but I 
think what we're debating is lineage of architectural evolution vs. a 
new processor that shares many (most?) features of an old processor to 
make programmers more comfortable.

   You and I are both clearly familiar with these components and their 
architectures.  I believe we should either agree to disagree on the 
lineage point (which may boil down to opinion anyway), or dissect the 
architectures feature by feature until one of us convinces the other.  
I've got my databooks handy. ;)

         -Dave

--
Dave McGuire             "I've grown hair again, just
St. Petersburg, FL           for the occasion."       -Doc Shipley