"Bruno Albuquerque" <bga@xxxxxxxxxxxxx> wrote: > On Thu, 23 Jul 2009 17:44:55 +0200 CEST, Axel Dörfler said: > Yep. but, as I said, the test was executed with the VM off. It does > improve things but the problem still happens. Could it be our page_ > writter holding a lock that is affecting memory allocations > everywhere? > This is just a guess, I did not look at the code yet but i am getting > there. This is a really annoying problem and does not look well in > demos. :) I'm more worried about the real life performance, but anyway, I guess Ingo and I could have a look at that problem next week. Bye, Axel.