----- Original Message ----- 
From: "Timothy Sipples" <[EMAIL PROTECTED]>
Newsgroups: bit.listserv.ibm-main
To: <[email protected]>
Sent: Sunday, July 15, 2007 11:54 PM
Subject: Re: Links to decent 'why the mainframe thrives' article


Dean Kent wrote:
>>I find it hard to believe
>>that IBM would spend the money for mainframe
>>processors to keep pace when there isn't really
>>much of an incentive to do so.

>Why is it hard to believe that IBM would invest huge sums in a fast growing
>market with significant demand elasticity?  That's logical and rational
>business behavior.

Because I wasn't talking about simply investing money in the platform.  I
was talking about investing money in the CPU development and manufacturing
process to keep pace in performance with x86.   The reason for the extremely
fast pace of improvement in the x86 world is the intense competition between
Intel and AMD, and the marketing based upon processor performance (vs.
system performance).  IBM doesn't really have that with the mainframe
market, so the incentive to spend a lot of effort in improving *processor*
performance isn't going to be as great.

And as for 'fast growing', I guess that's a relative term.   IBM's own
numbers (from the IT Jungle article) indicate 10,000 mainframe 'footprints'.
I'd heard previously that there were about 13.5K in the mainframe's heyday,
but the article says it may have been as many as 20,000.    Compare that
with PCs of all flavors, where there are more than a hundred million sold
every year, with most sold on performance claims.   The money in those
10,000 footprints is certainly significant, but much of it is software and
services, not the hardware itself.    I suspect that much of the $1.2B
investment in the mainframe is more about the system, the software and
related items than specific to improving the speed of the processor itself.

IBM likes to use MIPS as a measure of growth, but that is sort of like using
SPEC benchmark numbers to show how fast the x86 market is growing.   For
example, using a Dell Precision Workstation 330 system with a 1.4GHz P4 in
Nov 2000 the SPECint_rate was 5.80.   In June 2006, a Dell Precision
Workstation 390 with a 2.93GHz Intel Core Duo processor was 63.6.
Obviously one cannot make a direct comparison between MIPS and SPEC rate,
but the point is that both are trying to compare relative speed of the
system.   The IT Jungle article says that over 7 years there was a four-fold
increase in MIPS.    In a 6 year period, Intel improved their SPECint_rate
score by over 8 times.

Hence my statement that I find it "hard to believe IBM would spend the money
for mainframe processors to keep pace".    Context is important.

Regards,
   Dean

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to