[daip] [forwarded from Joanne M. Attridge] benchmark

[Patrick P. Murphy]

Faster and faster.  Just as well Eric's working on the DDT2K/Y2K...
------- start of forwarded message (RFC 934 encapsulation) -------
Content-Length: 1454
From: "Joanne M. Attridge" <jma at fritz.haystack.mit.edu>
To: pmurphy at NRAO.EDU
Subject: benchmark
Date: Fri, 9 Jun 2000 17:23:39 -0400 (EDT)
Message-ID: <Pine.LNX.4.10.10006091707240.4919-100000 at fritz.haystack.mit.edu>

Hi Pat,

I have a few new machines here and have been installing AIPS. Here's
a new benchmark for you...

SETPAR parameter 35 was set to 25.64
Dell - Intel Pentium III 750 MHz single processor
128 MB RAM
One 20 GB EIDE disk
Linux 2.2.14 (RedHat 6.2)
g77-egcs 2.91.66
AIPS version 31DEC99 (as of June 2nd, 2000)
Am(93)=27.78

/sbin/hdparm -c1 -d1 -m16 -u1 -X66 /dev/hda 

I couldn't do a "Flatters' prediction" because I couldn't find the
SPECfp(95) or SPECint(95) *anywhere* for the 750 MHz chip.

I also iterated the SETPAR parameter 35 as I ran the DDT to get the
best value.

I do know that the hdparm made a *huge* difference. If you are interested,
I will de-optimize the disk and run the DDTs again. I haven't yet run
the DDTs on the non-optimized system - I just know that POSSM runs MUCH
faster...

- --joanne

- -----------------------------------------------------------------------
 Dr. Joanne M. Attridge        |                                 
 MIT Haystack Observatory      |       Thought for the day:      
 Route 40                      |                                 
 Westford, MA 01886            | A bus station is where a bus stops.
 (978) 692-4764                | A train station is where a train stops.
 (781) 981-5193 (direct line)  | On my desk I have a workstation...
 (781) 981-0590 FAX            |
 jattridge at haystack.mit.edu    |                  
- ------------------------------------------------------------------------
------- end -------