Re: NUMCPUVPS and Fragments

Since one scan thread tends to be associated with one fragment, I would (for maximum performance anyway) go with lots of fragments, and let the thread dispatching algorithm sort 'em out. If the threads aren't there, the work can't get done, so load em up I say... Remember, you are doing mostly disk I/O, along with probably some index scanning, which really isn't much work, and one CPU can generally drive lots of disks to distraction, so for max parallelism, the more fragments the merrier. Enjoy dimming the lights...gm. Paul Drake wrote: > > On page 2-11 of the Informix Performance Guide Manual (Ver. 7.2), it states > that "queries perform best when the number of scan threads is a multiple > (or factor) of the number of CPU VPs". It then gives the example of having > 6 CPU VPs, and a table with 10 fragments, and suggesting to reduce the VPs > to 5. > > I was wondering if anyone could tell me, Why? snip......