Date: Fri, 22 Oct 2004 18:42:55 +0100 (BST) From: Robert Watson <rwatson@freebsd.org> To: Mike Tancsa <mike@sentex.net> Cc: freebsd-current@freebsd.org Subject: Re: device apic on a single processor machine Message-ID: <Pine.NEB.3.96L.1041022184042.55680C-100000@fledge.watson.org> In-Reply-To: <6.1.2.0.0.20041022113405.08fe2c48@64.7.153.2>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 22 Oct 2004, Mike Tancsa wrote: > When moving from RELENG_4 to RELENG_5, I noticed that in GENERIC, the > options > > options SMP # Symmetric MultiProcessor Kernel > device apic # I/O APIC > > are enabled by default. Going forward, is this the best thing to leave > in my default kernel on a uniprocessor machine ? I am not using the ULE > scheduler either and have hyperthreading disabled in the BIOS. > > I did a search on google, and in 2003 it was said not to having either > on a single processor machine but its not clear if this is no longer the > case. Right now, compiling SMP support into a kernel used on a UP-only system results in a substantial performance degredation because you pay the overhead of locking without the benefits of parallelism. There's a variety of work going on to hack away at aspects of that problem, but my general advice is to compile out SMP on UP systems to get maximum performance. I've been pushing for us to ship a UP kernel and an SMP kernel, and pick the UP by default, documenting how to get to SMP (similar to previous versions of FreeBSD) but it looks like this may not happen for 5.3. The logic being: most systems are still UP, so UP as a default provides the most benefit. It's also POLA from previous versions, and SMP users are more likely to be understanding about having to recompile to get best performance. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Principal Research Scientist, McAfee Research
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1041022184042.55680C-100000>