Date: Sat, 28 Jul 2007 19:03:54 -0300 From: JoaoBR <joao@matik.com.br> To: Kris Kennaway <kris@obsecurity.org> Cc: freebsd-stable@freebsd.org Subject: Re: powerd freeze with amd 5000 X2 but not with lower cpus Message-ID: <200707281903.54973.joao@matik.com.br> In-Reply-To: <20070728174717.GA66065@rot26.obsecurity.org> References: <200707271109.51334.joao@matik.com.br> <46AB48A7.8060103@freebsd.org> <20070728174717.GA66065@rot26.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday 28 July 2007 14:47:17 Kris Kennaway wrote: > On Sat, Jul 28, 2007 at 03:46:15PM +0200, Michael Nottebrock wrote: > > Meanwhile I found a workaround for my system: I had SCHED_ULE configured > > in my kernel - switching to SCHED_4BSD gets rid of the freezes. Should > > have thought of that sooner, ISTR having problems with powerd and > > SCHED_ULE even on single cpu P4s. > > I see you've now learned a reason why no-one should be using SCHED_ULE > on FreeBSD < 7 so using ULE in 7 is ok ? ? ? you already told once ULE is broken in 6 what is nonsense as probably using= =20 ULE in 7 is ... ULE in 6.x is absolutely ok and it runs depending on situation faster than= =20 4BSD with correct kernel and sysctl settings for it and it is perfectly=20 stable, specially with polling + net.isr.enable + net.inet.ip.forwarding += =20 some other tweaks depending on the servers load and several NICs on a route= r=20 probably up to 4-6MB continuous throughput when then 4BSD gets faster but n= ot=20 so much, small and midsize MySQL seems to be faster with ULE too especially= =20 with small r/w packages ULE also seems to be faster on a desktop with SMP and KDE on X2 CPUs and yo= u=20 can feel it=20 =2D-=20 Jo=E3o A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura. Service fornecido pelo Datacenter Matik https://datacenter.matik.com.br
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200707281903.54973.joao>