Subject: Big puzzle on processing times (WUs with same dates, AR & TerFlops)
From: John
Date: 31/12/2003, 17:18
Newsgroups: alt.sci.seti

I have just processed two couples of different WUs, and got radically
different processing times.

The first group of two WU is -

        25au03ab.6263.3617.573572.101 
        AR 0.547, TeraFlops 3.986, crunch time 5.054 hrs (about normal)

        25au03ab.6263.3617.573572.28
        AR 0.547, TeraFlops 3.984, crunch time 6.436 hrs (like a VLAR)

Second group of two WUs -

        25au03ab.6263.3617.573572.87
        AR 0.547, TeraFlops 3.991,  crunch time 5.233 hrs (about normal)

        25au03ab.6263.3617.573572.28
        AR 0.547, TeraFlops 3.991, crunch time 6.433 hrs (like a VLAR)
  
I have been using my PC box for word processing during both these sets
of results today.

Anyone able to speculate as to the wild crunching time differences?

NOTE:   I run 4 instances of the S@H CLI, under Driver & Spy to a proxy 
        (SETIQueue) on my box. It is a dual 2.8GHz Xeon, with HT active,
        under WinXP Pro. This box is part of a small farm.

        With this set up I achieve times for VLARs @ 6.0-6.5 hrs and for
        VHAR the time drops to 4.0-4.25 hrs. Normal AR processing is in
        the range 4.75-5.25 hrs.

 
-- John Clark