Date: Sat, 25 Apr 2015 10:44:16 -0700 From: Adrian Chadd <adrian@freebsd.org> To: "K. Macy" <kmacy@freebsd.org> Cc: "freebsd-arch@freebsd.org" <freebsd-arch@freebsd.org> Subject: Re: RFC: setting performance_cx_lowest=C2 in -HEAD to avoid lock contention on many-CPU boxes Message-ID: <CAJ-VmokhS7KLxq0fVRjDaR3naEMORei07=49SoiP2bSQXog8WQ@mail.gmail.com> In-Reply-To: <CAJ-Vmo=6ej6b52yoknP_piL5uJYHm_1AxZh7=6AFGuqMO4sLVg@mail.gmail.com> References: <CAJ-VmonG%2By5gzoYmer70KAswUorvezcZxRSDsQWj47=jsAZ71w@mail.gmail.com> <CAHM0Q_NuvZH0GGs-J9xniyt2PZ0qb_kjOaanVOaCxzD0CVzGYg@mail.gmail.com> <CAJ-Vmo=6ej6b52yoknP_piL5uJYHm_1AxZh7=6AFGuqMO4sLVg@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Oh the other thing, which I just mentioned to kip in IRC - all of the intel laptops I've tested (and that's a long list) don't enter CPU C7 if the power is plugged in. Ie: * power in, ACPI C2 -> CPU C6 * power in, ACPI C3 - CPU C6 * battery - ACPI C2 -> CPU C6 * battery -> ACPI C3 -> CPU C7 So having performance_cx_lowest=C3 is effectively a no-op on the devices that it'd matter on, so it's okay to just flip it to C2. -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmokhS7KLxq0fVRjDaR3naEMORei07=49SoiP2bSQXog8WQ>