Date: Thu, 18 Nov 2010 09:15:52 -0500 From: George Neville-Neil <gnn@neville-neil.com> To: Andriy Gapon <avg@FreeBSD.org> Cc: freebsd-acpi@freebsd.org, freebsd-current@freebsd.org Subject: Re: aperf/mperf Message-ID: <2A8C55CE-AD3A-419E-9028-EF8407596329@neville-neil.com> In-Reply-To: <4CE51CDA.6010202@freebsd.org> References: <4CE29718.2050508@freebsd.org> <D1DB20AD-779E-469B-BFFA-C0BA1A249858@neville-neil.com> <4CE51CDA.6010202@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Nov 18, 2010, at 07:32 , Andriy Gapon wrote: > on 18/11/2010 05:53 George Neville-Neil said the following: >>=20 >> On Nov 16, 2010, at 09:37 , Andriy Gapon wrote: >>=20 >>>=20 >>> Many modern processors provide APERF and MPERF MSRs which allow to = easily and >>> reliable calculate average CPU performance level over some interval = of time. >>> This also allows to notice things like performance boost, which is = generally >>> hidden from software. >>> What would be a proper place to add code that would measure = APERF/MPERF ratio? >>> When should trigger such a measurement and over what interval? >>> Ideas? >>=20 >> Can you point me at documentation for this? This sounds a lot like >> hwpmc(4) and I wonder if we can make these available in the same way. >=20 > Actually it feels more cpufreq-ish to me. > This feature is documented in, e.g., Intel Software Developer's Manual = volume 3A, > section 14.2 P-STATE HARDWARE COORDINATION. Ah, yes, quite right on cpufreq etc. Thanks for the documentation = pointer though. Best, George
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2A8C55CE-AD3A-419E-9028-EF8407596329>