On Mon, 29 Dec 2014 09:26:44 -0500, Shmuel Metz (Seymour J.) 
<[email protected]> wrote:
>That's not what you folks wrote when you introduced the term;
>technical articles described a hierarchy of microcode, millicode and
>z, with the millicode using an extended subset of the z instruction
>set and the microcode using an undocumented architecture, presumably
>VLIW (horizontal).

Yes, lots of articles, and at the end of the day, you still have no idea how 
the machine you have is implemented.  What runs where changes depending on the 
design target for the machine, ease of implementation, performance sensitivity, 
and things as mundane or esoteric as chip real estate or the date.

So while there is a difference in the execution environment of microcode and 
millicode, I don't find it a useful distinction except insofar that I know what 
engineering group to talk to about something.  I have seen programmers get 
burned by making rash assumptions about implementation and trying to 
micromanage (or outwit?) the microcode, so to speak.

Don't get me wrong.  A machine that was designed to let you change the 
underlying hardware design without altering the programming architecture of the 
machine was very smart.  And way cool.  And a major move forward in the 
industry.   There's a reason the architecture has survived for 50 years with 
those old 24-bit applications still running just fine thank-you-very-much.

But my experience within IBM is that that we love talking about technology.  In 
fact, we sometime love it a bit too much, particularly when it's new and it 
hasn't completed its evolutionary journey.

Alan Altmark
IBM

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to