HP3000-L Archives

September 2002, Week 1

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jeff Woods <[log in to unmask]>
Reply To:
Date:
Thu, 5 Sep 2002 13:00:57 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (41 lines)
Gavin wrote:
>All current A-class and some N-class systems are "clock
>reduced" meaning they have some sort of delay loop in MPE that
>prevents you from using all of the CPU cycles of the
>(typically) 440MHz processor in the box.

And it appears that it also prevents you from using all the I/O and
memory bus cycles at the same time.

>As a result of doing MPE to HP-UX performance comparisons it
>has recently been determined that the "110MHz" numbers appear
>to have been made up by HP marketing, and that these boxes
>actually only allow you to use about 55MHz out of the 440MHz
>available, so the "good news" is that your "110MHz" A-class
>MPE system will actually become about *eight* times faster in
>CPU speed when turned into an uncrippled HP-UX system.

So this means an "110MHz" (really 50-something MHz) A400 is throwing
away 7/8 of the whole machines throughput.  Can I have my daily
allotment of 3 hours processing throughput all at once?  We could do
daily production in the first hour, evening batch processing in the
second hour and "overnight backup" in the third hour... All in the
course of a short afternoon!

P.S.  In case it's not obvious, I personally find artificial hardware
performance crippling to be extremely distasteful.  Fortunately, Linux,
BSD and other open-source environments aren't conducive to such
"marketing factors".

P.P.S.  The spell-checker in this email client complains that
"uncrippled" isn't a word.  Is the correct spelling "repaired"?
--
Jeff Woods
[log in to unmask]
#pragma standard_disclaimer
/* I don't speak for my employer even if they happen to agree with me.
*/

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2