Martin 53N 1W wrote:

 From the days of old of s@h classic, I have old and dusty recollections
that the time to complete a WU depended on the Angle Range of that work
unit.

Anyone have any ideas /why/?

What was *different about the processing* ?
And how did this make some systems execute faster for a particular AR
whilst others would be slower?

Regards,
Martin

http://www.teamlambchop.com/runtime303.htm

http://www.setiteam.co.uk/print.php?sid=15

"The amount of time taken to process a work unit can vary quite a lot from
one work unit to the next, and it's all dependant on the type of data
within the unit.

VLARs and VHARs

A VLAR (very low angle range) work unit is usually the most difficult to
process. In particular the windows client has a hard time dealing with
these VLARs whereas other clients (such as Linux) will manage them better.
These units on windows clients can take up to one and a half times as long
to deal with than 'normal' units.

A VHAR (very high angle range) work unit (also known as a sweet unit) are
the fastest units to process, generally they can be processed around twenty
percent faster than normal units.

Avoiding VLARs

Using a program such as SETI Queue to sending and receive your results can
help you to avoid the difficult work units. SETI Queue can be set to try an
obtain as many sweet units as it can, it does this by indentifying any work
unit that is good and immediately requesting more units which will usually
also be sweet.

SETI Queue can also be setup to try and route the more difficult (VLARs)
units to linux machines, which cope with them far better.

For more information on processing times and angle ranges see this excellent
article , and this one, and it's follow up over at the Team Lambchop site."