Date: Sat, 04 Feb 2006 08:59:17 -0700 (MST) From: "M. Warner Losh" <imp@bsdimp.com> To: ivoras@fer.hr Cc: gallatin@cs.duke.edu, current@freebsd.org Subject: Re: machdep.cpu_idle_hlt and SMP perf? Message-ID: <20060204.085917.129781502.imp@bsdimp.com> In-Reply-To: <43E45BD8.7070708@fer.hr> References: <17379.56708.421007.613310@grasshopper.cs.duke.edu> <43E45BD8.7070708@fer.hr>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <43E45BD8.7070708@fer.hr>
Ivan Voras <ivoras@fer.hr> writes:
: Andrew Gallatin wrote:
: > Why dooes machdep.cpu_idle_hlt=1 drop my 10GbE network rx
: > performance by a considerable amount (7.5Gbs -> 5.5Gbs)?
:
: For what it's worth, here are results of unixbench's "context1"
: benchmark on an old P3 SMP with and without cpu_idle_hlt (6-stable):
When ACPI CPU states were coming into the tree, there was discussion
that they had been invented to provide a lower energy idle as well as
a faster transition from the idle state to something else. I think
that these benchmarks show this effect rather dramatically.
Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060204.085917.129781502.imp>