X-no-Archive:yes
This is why I have asked about TPC benchmarks, and have not yet given up on
seeing what I can do with TPC-C in all my spare time ;-). Instead of
comparing apples to oranges, you can compare results to cost. Then it's
apples and oranges again, why one number at one cost is better than another
number at a different cost. And sure, the database engines are all very
different from each other. But at least there is a hard number according to
someone besides just the vendor, audited by a third-party, that lets us make
certain comparisons.

Greg Stigers
http://www.cgiusa.com

-----Original Message-----
From: John Painter [mailto:[log in to unmask]]
Sent: Thursday, August 10, 2000 12:37 PM
To: [log in to unmask]
Subject: Re: OLTP stats for 3000s

John:

There is no such animal. HP doesn't benchmark the 3000 with standard
benchmarks like TPC, etc. Plus, most folks are running IMAGE, so there
is no database comparison that can be easily made. Even though an older
Oracle version runs on the 3K, I don't know of any performance
comparison there either. You used to be able to match up some 9000
models with 3000 models, but that didn't tell you much-- it would be
like comparing a PC running Linux with a PC running Windows. The opsys
is different, so it's really apples and oranges without a standardized
benchmark. The 3k has always been sold on the recommendation of a
specific software vendor who said: this is the best platform for my
application. If that isn't good enough, the customer ends up somewhere
else.

If you want a chart comparing just 3000s to each other-- you can find
one on our web site under "downloads".

John Painter
Computer Solutions, Inc.
http://www.internetcsi.com