>The dynamic SQL is 2 times faster than the SP. Can you guys confirm this?

The benchmark rings a bell, but I have no link / reference at the moment. AFAIR the gist was, that on identical dynamic queries (at least) the whole execution plan is cached, whereas with SP the parameter is always "inserted" - but they wanted to optimize further. So perhaps a different [better?] test would be to have lnPercentage change inside the loop. And since when are you mDotting the left side of assignments as in

m.lnPercentage = 30
m.lnElapsed = seconds() - m.a

??

regards

thomas