In article <f4hB5JA6Tw8$EwFn@consltec.demon.co.uk>, John <fredclark@cons
ltec.demon.co.uk> writes

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.

Whether it is relevant or not (I don't think so) the PC has not been
rebooted for over a month now.
-- 
John Clark