Postgresql on multi-core CPU's: is this old news?
От | Mischa Sandberg |
---|---|
Тема | Postgresql on multi-core CPU's: is this old news? |
Дата | |
Msg-id | CB0F73E165CFFA4496D12161D835562C030AE939BA@US-COL-EXCHMBX1.green.sophos обсуждение исходный текст |
Ответы |
Re: Postgresql on multi-core CPU's: is this old news?
Re: Postgresql on multi-core CPU's: is this old news? |
Список | pgsql-hackers |
<div class="WordSection1"><p class="MsoNormal">Came across the following in a paper from Oct 2010. Was wondering is thisis old news I missed in this group.<p class="MsoNormal"><a href="http://pdos.csail.mit.edu/papers/linux:osdi10.pdf">http://pdos.csail.mit.edu/papers/linux:osdi10.pdf</a><p class="MsoNormal">aboutLinux optimization on multi-core CPU’s.<p class="MsoNormal"> <p class="MsoNormal">The group at MITwere exploring how some Linux apps were scaling up --- sometimes badly, mostly due to hidden contention over cache-lineconsistency across the cores’ caches.<p class="MsoNormal">In a nutshell: if an app, or the system calls it uses,tries to modify anything in a cache line (32-64 byte slice of memory) that another core is using, there’s a lot of fumblingin the dark to make sure there is no conflict. When I saw PostgreSQL named in the abstract, I thought, “Aha! Contentionover shm”. Not so. Skip to page 11 (section 5.5) for most of the PG specifics.</div>
В списке pgsql-hackers по дате отправления: