From: Dirk Brandewie <dirk.brandewie@gmail.com>
To: lkp@lists.01.org
Subject: Re: [cpufreq/intel_pstate] 1abc4b20b85: +7.3% turbostat.Pkg_W
Date: Wed, 19 Feb 2014 07:46:44 -0800 [thread overview]
Message-ID: <5304D1E4.2070009@intel.com> (raw)
In-Reply-To: <20140219140628.GA8097@localhost>
[-- Attachment #1: Type: text/plain, Size: 2523 bytes --]
On 02/19/2014 06:06 AM, Fengguang Wu wrote:
> Hi Dirk,
>
> FYI, we noticed the below changes on commit 1abc4b20b85b42e8573957e54b193385cf48b0d6
> ("cpufreq / intel_pstate: remove idle time and duration from sample and calculations"):
>
This commit was a LONG time ago. How does this compare to the current state of
the driver?
This functionality was added back in commit fcb6a15c2e7e76
--Dirk
> test case: snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqrd-sync
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 1.61 ~ 0% +32.2% 2.13 ~ 0% TOTAL turbostat.GHz
> 0.86 ~ 3% +15.3% 0.99 ~ 1% TOTAL turbostat.%pc3
> 11.03 ~ 1% -11.9% 9.72 ~ 1% TOTAL turbostat.%c1
> 5.36 ~ 0% -21.8% 4.19 ~ 0% TOTAL turbostat.%c0
> 2.67 ~ 0% +18.2% 3.15 ~ 0% TOTAL turbostat.Cor_W
> 6.66 ~ 0% +7.3% 7.15 ~ 0% TOTAL turbostat.Pkg_W
> 1278688 ~ 4% -16.6% 1066682 ~ 4% TOTAL cpuidle.C1-SNB.usage
> 70490 ~ 1% -13.1% 61235 ~ 1% TOTAL cpuidle.C3-SNB.usage
> 472542 ~ 2% -13.7% 408029 ~ 2% TOTAL cpuidle.C1E-SNB.usage
> 1.788e+08 ~ 5% -15.4% 1.512e+08 ~ 5% TOTAL cpuidle.C1-SNB.time
> 3.777e+08 ~ 1% -12.4% 3.311e+08 ~ 0% TOTAL cpuidle.C1E-SNB.time
> 2213808 ~ 1% -14.6% 1890218 ~ 0% TOTAL interrupts.RES
> 920130 ~ 5% -13.9% 791994 ~ 5% TOTAL interrupts.LOC
> 52713 ~ 0% -14.9% 44878 ~ 0% TOTAL softirqs.TIMER
>
> test case: hackbench, netperf and dbench
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 312046 ~ 1% -47.3% 164399 ~ 0% lkp-snb01/micro/hackbench/1600%-threads-pipe
> 151815 ~ 0% -0.8% 150593 ~ 0% lkp-snb01/micro/hackbench/1600%-threads-socket
> 463862 ~ 0% -32.1% 314993 ~ 0% TOTAL hackbench.throughput
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 2761 ~ 0% -1.8% 2713 ~ 0% lkp-a04/micro/netperf/120s-200%-UDP_RR
> 2761 ~ 0% -1.8% 2713 ~ 0% TOTAL netperf.Throughput_tps
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 8276 ~ 0% -0.3% 8247 ~ 0% nhm8/micro/dbench/100%
> 8276 ~ 0% -0.3% 8247 ~ 0% TOTAL dbench.throughput-MB/sec
>
> Thanks,
> Fengguang
>
WARNING: multiple messages have this Message-ID (diff)
From: Dirk Brandewie <dirk.brandewie@gmail.com>
To: Fengguang Wu <fengguang.wu@intel.com>
Cc: dirk.j.brandewie@intel.com,
"Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
LKML <linux-kernel@vger.kernel.org>,
lkp@01.org
Subject: Re: [cpufreq/intel_pstate] 1abc4b20b85: +7.3% turbostat.Pkg_W
Date: Wed, 19 Feb 2014 07:46:44 -0800 [thread overview]
Message-ID: <5304D1E4.2070009@intel.com> (raw)
In-Reply-To: <20140219140628.GA8097@localhost>
On 02/19/2014 06:06 AM, Fengguang Wu wrote:
> Hi Dirk,
>
> FYI, we noticed the below changes on commit 1abc4b20b85b42e8573957e54b193385cf48b0d6
> ("cpufreq / intel_pstate: remove idle time and duration from sample and calculations"):
>
This commit was a LONG time ago. How does this compare to the current state of
the driver?
This functionality was added back in commit fcb6a15c2e7e76
--Dirk
> test case: snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqrd-sync
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 1.61 ~ 0% +32.2% 2.13 ~ 0% TOTAL turbostat.GHz
> 0.86 ~ 3% +15.3% 0.99 ~ 1% TOTAL turbostat.%pc3
> 11.03 ~ 1% -11.9% 9.72 ~ 1% TOTAL turbostat.%c1
> 5.36 ~ 0% -21.8% 4.19 ~ 0% TOTAL turbostat.%c0
> 2.67 ~ 0% +18.2% 3.15 ~ 0% TOTAL turbostat.Cor_W
> 6.66 ~ 0% +7.3% 7.15 ~ 0% TOTAL turbostat.Pkg_W
> 1278688 ~ 4% -16.6% 1066682 ~ 4% TOTAL cpuidle.C1-SNB.usage
> 70490 ~ 1% -13.1% 61235 ~ 1% TOTAL cpuidle.C3-SNB.usage
> 472542 ~ 2% -13.7% 408029 ~ 2% TOTAL cpuidle.C1E-SNB.usage
> 1.788e+08 ~ 5% -15.4% 1.512e+08 ~ 5% TOTAL cpuidle.C1-SNB.time
> 3.777e+08 ~ 1% -12.4% 3.311e+08 ~ 0% TOTAL cpuidle.C1E-SNB.time
> 2213808 ~ 1% -14.6% 1890218 ~ 0% TOTAL interrupts.RES
> 920130 ~ 5% -13.9% 791994 ~ 5% TOTAL interrupts.LOC
> 52713 ~ 0% -14.9% 44878 ~ 0% TOTAL softirqs.TIMER
>
> test case: hackbench, netperf and dbench
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 312046 ~ 1% -47.3% 164399 ~ 0% lkp-snb01/micro/hackbench/1600%-threads-pipe
> 151815 ~ 0% -0.8% 150593 ~ 0% lkp-snb01/micro/hackbench/1600%-threads-socket
> 463862 ~ 0% -32.1% 314993 ~ 0% TOTAL hackbench.throughput
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 2761 ~ 0% -1.8% 2713 ~ 0% lkp-a04/micro/netperf/120s-200%-UDP_RR
> 2761 ~ 0% -1.8% 2713 ~ 0% TOTAL netperf.Throughput_tps
>
> 559f56c70fc90bd 1abc4b20b85b42e8573957e54
> --------------- -------------------------
> 8276 ~ 0% -0.3% 8247 ~ 0% nhm8/micro/dbench/100%
> 8276 ~ 0% -0.3% 8247 ~ 0% TOTAL dbench.throughput-MB/sec
>
> Thanks,
> Fengguang
>
next prev parent reply other threads:[~2014-02-19 15:46 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-02-19 14:06 [cpufreq/intel_pstate] 1abc4b20b85: +7.3% turbostat.Pkg_W Fengguang Wu
2014-02-19 14:06 ` Fengguang Wu
2014-02-19 15:46 ` Dirk Brandewie [this message]
2014-02-19 15:46 ` Dirk Brandewie
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5304D1E4.2070009@intel.com \
--to=dirk.brandewie@gmail.com \
--cc=lkp@lists.01.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.