David,
You have discovered one of the reasons why I have a job at HP! If it was as
simple as measuring the MHz ratings of the CPU, we would not need to do any
performance benchmarking on the HP e3000 family. But we do need to!
Real customer applications do a lot more than just use raw CPU on a 3000.
They involve memory, disk IO, networking, and database locking strategies,
to name a few.
The 3.2 number measured for a A500-100-140 was arrived at by measuring real
customer applications on that box, and comparing them to the results for
earlier systems. The comparable number for the original 987 is 4.2. The
numbers quoted below by another poster are based on an older scale we used
in the past, and are not comparable with the 3.2 for the A500-100-140.
I gave a paper at HPWorld this year that included info on upgrade paths,
and the MINIMUM suggested upgrade for a 987/100 is to the A500-200-140,
which is rated at 5.4. To see the slides from this presentation, see:
www.hp.com/products1/mpeixservers/news_events/hpworld2001 and click on
the fourth paper in the list.
Also, I feel I need to mention that there were also two later versions
of the 987 - the 987/150 and the 987/200. You may be well aware of which
one you have, but I mention it just in case. Your upgrade path would be
different, depending on which one you have now. (See my HPWorld paper.)
And while the potential IO speeds of the A500 can reach 440MB/sec, there
are no disks available yet that can take advantage of those speeds on the
A500. The limiters are in the SCSI interfaces.
I hope that helps!
Kevin Cooper
HP e3000 Performance Team
[log in to unmask]
Joseph Dolliver ([log in to unmask]) wrote:
: HP 3000 9x7 Series
: System: 917 927 937 947 957 967 977 987 987/150 987/200
: Re 918: 1.0 1.0 1.0 1.0 1.6 2.0 2.6 3.2 4.5 6.0
: -----Original Message-----
: To the list,
: I'm thinking about getting rid of my 987 and moving to one of the new
: A-class servers. However, according to documentation I've found on the web,
: the 987 is considered a performance factor of about 4.2 and the A500 single
: CPU is 3.2. It looks like I'm going backwards. How is that possible?
: --The CPU is faster (96mhz compared to 140mhz)
: --the memory is faster (isn't it?) and there would be more (512meg compared
: to 1gig)
: --the IO would be faster (32MB/sec compared to 440MB/sec)
: Are the different relative performance factors based on expandability as
: well as raw performance?
: I've been trying to get help from HP, but I still can't seem to find anyone
: in the sales organization that knows anything about HP3000's...
: David Knispel
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|