I really hate to bring this to you speed, but here is the deal.
Seti client and seti driver both report CPU time per WU, not the actual time it takes to
finish a WU. In your pic, and your post seti is reporting total time of 14 hours. This is
14 CPU hours. Since the folding client is using approx 50% of your CPU time Seti will finish
one WU in 2x14 hours=28 hours.
Then in 28 hours your machine is pumping out 2 seti WU (one for each CPU) and if we assume 2:50 per
folding WU (like you say, and the pic shows) 28 hours=1680min / 170min this gives 9,88 WU per cpu
per 28 hours, and since you are running two =19,76 WU
Therefore in 28 hours = 2x seti and 19,76 folding WU.
Alternative#1
Two seti @ 11 h/wu each, in 28 hours = 2x(28/11) 5.1 WU
Alternative#2
Two folding @ 1,5 h/wu hours in 28 hours = 37,33 WU per 28 hours
Alternative#3
one seti
@11 h/wu (cpu1) = 2,56 WU per 28 hours
one folding @1,5 h/wu (cpu2) = 18,66 WU per 28 hours
You can see how similar alternative#3 is to the one on top, acually I am using conservative
folding times in #3 1,5 hours instead of the half of 2:50 for the folding WU time.
IMO you are wasting half seti WU per 28 hours by running this confi, this is the time lost
due to more overhead.
If you want to experiment you should try to run 4 instances of seti on your machine, two
on each CPU (without folding). You will see similar times per WU reported by seti, but
this does of course not increse the net wu output since this is the relative cpu time.
Happy crunching