Re: [HACKERS] Visibillity testing - some numbers on current performance.

2011-04-05 Thread Robert Haas
On Tue, Apr 5, 2011 at 3:04 PM, Jesper Krogh wrote: > I initially set out to put some numbers on "why" the visibillity > map was important for "select count(*)", primarily to give some > feedback to Simon Riggs stating: > "Your tests and discussion remind me that I haven't yet seen any tests > tha

[HACKERS] Visibillity testing - some numbers on current performance.

2011-04-05 Thread Jesper Krogh
Hi. I initially set out to put some numbers on "why" the visibillity map was important for "select count(*)", primarily to give some feedback to Simon Riggs stating: "Your tests and discussion remind me that I haven't yet seen any tests that show that index-only scans would be useful for performa