Message-ID: <1389686473.5951.11.camel@marge.simpson.net>
Subject: Re: Idle power fix regresses ebizzy performance (was 3.12-stable
 backport of NUMA balancing patches)
From: Mike Galbraith <bitbucket@online.de>
To: Len Brown <lenb@kernel.org>
Cc: Mel Gorman <mgorman@suse.de>, Len Brown <len.brown@intel.com>,
        athorlton@sgi.com, Rik van Riel <riel@redhat.com>, chegu_vinod@hp.com,
        Greg KH <gregkh@linuxfoundation.org>,
        "H. Peter Anvin" <hpa@linux.intel.com>,
        LKML <linux-kernel@vger.kernel.org>, stable@vger.kernel.org
Date: Tue, 14 Jan 2014 09:01:13 +0100
In-Reply-To: <CAJvTdKniGkxxzvPvEDSU9BgtubR-5Fn3YsepjSNAwn5XkL3nVA@mail.gmail.com>
References: <1389103248-17617-1-git-send-email-mgorman@suse.de>
	 <20140107141715.GA32491@kroah.com> <20140107185440.GA7844@kroah.com>
	 <20140107203012.GA27046@suse.de> <20140108104340.GC27046@suse.de>
	 <20140108134858.GF27046@suse.de> <20140113192406.GL27046@suse.de>
	 <CAJvTdKniGkxxzvPvEDSU9BgtubR-5Fn3YsepjSNAwn5XkL3nVA@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org

On Tue, 2014-01-14 at 02:31 -0500, Len Brown wrote: 
> > This is a false alarm.
> 
> Thanks for the follow-up, Mel.
> 
> Agreed, it makes no sense for ebizzy measure 'throughput', when a
> library debug bottleneck
> prevents it from scaling past 3% CPU utilization.
> 
> Still, the broken configuration did find a difference due to the
> addition of CLFLUSH on this box.
> It makes me wonder if we will find issues on workloads that may depend
> on the latency
> of idle entry/exit, or perhaps sensitivity to the state of the cache
> line containing thread_info->flags.
> 
> If somebody runs into such a workload, please try changing this 1 line
> of intel_idle.c to limit
> the CLFLUSH to C-states deeper than C1E, and let me know what you see.
> 
> - if (this_cpu_has(X86_FEATURE_CLFLUSH_MONITOR))
> + if ((eax > 1) && this_cpu_has(X86_FEATURE_CLFLUSH_MONITOR))
>           clflush((void *)&current_thread_info()->flags);

Hm, seems any high frequency switcher scheduling cross-core (pipe-test,
or maybe a tbench pair) should show the cost to an affected box.

-Mike

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/