Re: [HACKERS] Patch attached...
От | Don Baccus |
---|---|
Тема | Re: [HACKERS] Patch attached... |
Дата | |
Msg-id | 3.0.1.32.20000204203823.01030d30@mail.pacifier.com обсуждение исходный текст |
Ответ на | Patch attached... (Chris <chris@bitmead.com>) |
Ответы |
Re: [HACKERS] Patch attached...
|
Список | pgsql-hackers |
At 02:36 PM 2/5/00 +1100, Chris wrote: >*) The overhead for non-inheritance has >been cut down to 30 microseconds (on a pc). What kind of PC? I'm getting 4,000 microseconds doing simple selects on a classic P200 (no L2 cache) through AOLserver and Tcl scripts, which probably means more like 2,000 microseconds for PG alone. But without knowing your PC, I have no way to scale. For instance, my P500e that I just built gets between 3-6x performance over my P200. What's an acceptable level for overhead? I have no personal desire to eat any overhead, in all honesty. 2000/30 < 1% but without knowledge of the actual PC platform (which certainly you must know vary widely in performance) I have no way to scale. If your PC platform is closer to my P500e than my (classic) P200 (not pro, no L2 cache) then the overhead is more like 2-3%. That's measurable. And if SQL92 compliance is the goal, why must ANY degradation of performance be acceptable unless there are very, very strong reasons to do so (reasons that impact the target audience). - Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert Serviceand other goodies at http://donb.photo.net.
В списке pgsql-hackers по дате отправления: