From: Philippe Longepe <philippe.longepe@linux.intel.com>
To: linux-pm@vger.kernel.org
Cc: srinivas.pandruvada@linux.intel.com,
Philippe Longepe <philippe.longepe@intel.com>
Subject: [PATCH v1 0/5] intel_pstate: Use the cpu load first on Atom only
Date: Mon, 23 Nov 2015 14:06:35 +0100 [thread overview]
Message-ID: <1448284006-13596-1-git-send-email-philippe.longepe@linux.intel.com> (raw)
From: Philippe Longepe <philippe.longepe@intel.com>
This patch serie aims to use the load as the input of the PID.
The load is defined by cpu_load = 100 * (delta_mperf / delta_tsc)
since mperf is counting at HFM during C0 state and TSC is counting at
the same frequency during all C-States.
The load gives very good power improvements on Android manly for gaming
use cases.
Also include the IOBoost improvement that converts time spent in iowait
into mperf cycles to avoid performance degradation during IOs use cases.
Measured on a Cherrytrail with this algorithm:
Power (in J) Power with patch Power Gain
Candycrush Saga (30sec) 46.7 14.5 221.6%
Circular ProgressBar 10.4 3.2 226.12%
SmartBench 76 54.3 40%
FishTank 226 116 95%
AnTuTu5.6 315 311 2.7%
Philippe Longepe (5):
cpufreq: intel_pstate: change function name for calculation for busy
cpufreq: intel_pstate: Rename current busy calculation function
cpufreq: intel_pstate: account for non C0 time
cpufreq: intel_pstate: configurable busy calculation
cpufreq: intel_pstate: try load instead of busy_scaled
drivers/cpufreq/intel_pstate.c | 56 ++++++++++++++++++++++++++++++++++++------
1 file changed, 48 insertions(+), 8 deletions(-)
--
1.9.1
next reply other threads:[~2015-11-23 13:06 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-11-23 13:06 Philippe Longepe [this message]
2015-11-23 13:06 ` [PATCH v1 0/5] intel_pstate: Use the cpu load first on Atom only Philippe Longepe
2015-11-23 13:06 ` [PATCH v1 1/5] cpufreq: intel_pstate: change function name for calculation for busy Philippe Longepe
2015-11-23 13:06 ` [PATCH " Philippe Longepe
2015-11-23 13:06 ` [PATCH v1 2/5] cpufreq: intel_pstate: Rename current busy calculation function Philippe Longepe
2015-11-23 13:06 ` Philippe Longepe
2015-11-23 13:06 ` [PATCH v1 3/5] cpufreq: intel_pstate: account for non C0 time Philippe Longepe
2015-11-23 13:06 ` Philippe Longepe
2015-11-23 13:06 ` [PATCH v1 4/5] cpufreq: intel_pstate: configurable busy calculation Philippe Longepe
2015-11-23 13:06 ` Philippe Longepe
2015-11-23 13:06 ` [PATCH v1 5/5] cpufreq: intel_pstate: try load instead of busy_scaled Philippe Longepe
2015-11-23 16:10 ` Doug Smythies
2015-11-23 16:26 ` Philippe Longepe
2015-11-24 1:33 ` Doug Smythies
2015-11-23 13:06 ` Philippe Longepe
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=1448284006-13596-1-git-send-email-philippe.longepe@linux.intel.com \
--to=philippe.longepe@linux.intel.com \
--cc=linux-pm@vger.kernel.org \
--cc=philippe.longepe@intel.com \
--cc=srinivas.pandruvada@linux.intel.com \
/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.