From owner-freebsd-ppc@FreeBSD.ORG  Sat Mar 29 16:04:58 2014
Return-Path: <owner-freebsd-ppc@FreeBSD.ORG>
Delivered-To: ppc@freebsd.org
Received: by hub.freebsd.org (Postfix, from userid 1033)
 id 67ECE61A; Sat, 29 Mar 2014 16:04:58 +0000 (UTC)
Date: Sat, 29 Mar 2014 16:04:58 +0000
From: Alexey Dokuchaev <danfe@FreeBSD.org>
To: Joe Nosay <superbisquit@gmail.com>
Subject: Re: System clock falls behind quickly on Mac mini G4
Message-ID: <20140329160458.GA95482@FreeBSD.org>
References: <20140328071714.GA45961@regency.nsu.ru>
 <CAFY7cWBCFmtx4Tsg3=mSJyscpk5nCY3S6Sxy52TKEoTmy1sFPA@mail.gmail.com>
 <CA+WntOtEoqG_RJ2D9vSb7mO-UfT13RZbAb04p6rV2mWLBu=H9Q@mail.gmail.com>
 <20140329100134.GA7863@regency.nsu.ru>
 <CA+WntOup7_98LoCLJ43xWc3magTArKNVUEHg_2v1UFWoGp=HJQ@mail.gmail.com>
 <20140329130507.GA28559@FreeBSD.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20140329130507.GA28559@FreeBSD.org>
User-Agent: Mutt/1.5.22 (2013-10-16)
Cc: Alexey Dokuchaev <danfe@nsu.ru>, ppc@freebsd.org
X-BeenThere: freebsd-ppc@freebsd.org
X-Mailman-Version: 2.1.17
Precedence: list
List-Id: Porting FreeBSD to the PowerPC <freebsd-ppc.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/options/freebsd-ppc>,
 <mailto:freebsd-ppc-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-ppc/>
List-Post: <mailto:freebsd-ppc@freebsd.org>
List-Help: <mailto:freebsd-ppc-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-ppc>,
 <mailto:freebsd-ppc-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Sat, 29 Mar 2014 16:04:58 -0000

On Sat, Mar 29, 2014 at 01:05:07PM +0000, Alexey Dokuchaev wrote:
> On Sat, Mar 29, 2014 at 09:01:38AM -0400, Joe Nosay wrote:
> > I found from building on a G3 400MHz, a G4 800MHz, and a G4 933MHz that
> > increasing the clock rate to that value improved performance and decreased
> > lag. Try and see for yourself. Even 1500 works better than 1000. My opinion
> > based on observation.
> 
> Thanks for suggestion; I'll try.

I've done a "time make buildkernel -DNO_MODULES" (with GENERIC config) both
against kern.hz=1000 (default) and kern.hz=2000:

  kern.hz=1000: 1103.7u, 187.0s
  kern.hz=2000: 1108.1u, 168.7s

So it doesn't look like it plays a big role.

I'll stick with kern.hz=2000 for a while to see if it makes any difference
in interactive workloads though.

./danfe