HP3000-L Archives

August 1999, 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:
Wirt Atmar <[log in to unmask]>
Reply To:
Date:
Thu, 5 Aug 1999 14:48:37 EDT
Content-Type:
text/plain
Parts/Attachments:
text/plain (32 lines)
Gavin writes:

> HP9000 K420 (mostly full time) and Dell 400Mhz PII running Linux (at
>  night, weekends).  Both are almost exactly the same speed at ~10 hours
>  per work unit.  The K420 has two CPUs and runs two instances of the
>  SETI client, so the box produces two results per 10 hours on average.

>  Basically the Windows version of the client is terrible (or at least the
>  older ones were, I haven't run the recent ones).  The Dell PC is around
>  four times faster running the non-gui Linux SETI client than it is running
>  the Windows version.

Actually, I don't believe the difference between the Windows and Linux
versions are as great as Gavin suggests. I have only one PC that's comparable
to Gavin's Dell PII 400MHz Linux box. It's a Dell Celeron 400MHz Win98
machine. The time to calculate one work unit is about 11-12 hours, only
slightly slower than Gavin is reporting, even though the Celeron processor is
a significantly crippled version of the Pentium II.

Our reported times are much longer than 11 hours because we have three 200MHz
Pentium machines running most of the time -- and they are averaging about 25
hours a piece (as you would theoretically expect). The faster machines that
we own are generally in use now 18 to 20 hours a day so they simply don't get
much time to be devoted to the SETI calculations.

On Windows, it is important to set your machine to go to a blank screen saver
after just a few minutes. The SETI client's graphics otherwise present a
substantial load to the PC. By turning the screen saver to blank, you will
speed up your work unit calculations by about a factor of 4.

Wirt Atmar

ATOM RSS1 RSS2