At 04:48 PM 2/9/2003 -0500, Leyne, Sean wrote:

>Jim,
>
>> If this is indeed the case, a simple dedicated page write thread
>> may give a performance boost when forced write is enabled.
>
>Do you think this would benefit both SuperServer and Classic engines?
>
>(I think more likely SuperServer -- watch I'll be wrong)
>

If I remember correctly, the engine "releases" the de facto
mutex during I/O, allowing another thread the run. In SuperServer,
there's probably a thread waiting. In Classic, probably not.

I fundamentally don't believe in performance predictions by
analysis -- there are usually enough complications and interactions
that big surprises are common. I've always favored the "code
the sucker and measure it" approach.

Jim Starkey