Hi,

Multi level cache, translation lookaside buffers and a number of other items
come to mind.

Paul

-----Original Message-----
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Phil
Payne
Sent: Saturday, December 13, 2003 5:30 AM
To: [EMAIL PROTECTED]
Subject: Re: Technical Specs


> And in each generation the pipelining abilities get better and better.

Not always.  3033 ---> 3081?

> For those who aren't up on the latest in processor design, 
> "pipelining" is a way for the CPU to actually execute more than one 
> instruction (or parts of a multi-cycle instruction) a single clock 
> cycle.  Optimizing compilers can build instruction streams suited to 
> the capabilities of a particular processor, throwing the relationship 
> between instruction execution and cycle time (MHz) out the window.  
> You could end up with a 1GHz CPU executing instructions as though it 
> were, say, 1.3GHz.  (I'm making those numbers up.)

A simplification.  The hardware does a whole RAFT of things that serve more
to speed up the _environment_ than to speed up instruction execution.  One
major difference between the Intel environment and the IBM mainframe
environment is that the former is principally designed for a single user and
is therefore not strongly optimised for support of high-intensity task
switching.  The mainframe is; and so there's a whole load of hardware
designed to preserve and restore the environment of an interrupted task as
efficiently as possible.

Run a single instruction loop and a modern Intel processor will clean the
clock of even a z990.  No contest.

Start some serious task/address space switching and watch the ratios change
- z990, at less than half the clock rate, will just leave Intel's best in
the dust.

--
  Phil Payne
  http://www.isham-research.com
  +44 7785 302 803

Reply via email to