Qingqing Zhou wrote:

"Bruce Momjian" <pgman@candle.pha.pa.us> wrote
Added to TODO list.
One thing we tried in February was padding out the statically defined
locks with dummy lock definitions in the enum. This has the effect of
ensuring that the most contested locks are very definitely in their own
cache line and not shared with others.
That showed a noticeable improvement in performance, probably because it
costs very little to implement, even if the code would require some
explanatory documentation.
Has this been done? See the LWLOCK_PADDED_SIZE macro in code.
Oh, yes, thanks. I thought it had but I couldn't find anything in the
area of the code he propsed the patch.

--
Bruce Momjian http://candle.pha.pa.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

Search Discussions

Discussion Posts

Previous

Follow ups

Related Discussions

People

Translate

site design / logo © 2017 Grokbase