lkml.org 
[lkml]   [2008]   [Dec]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Subject[PATCH v7 0/8] Tunable sched_mc_power_savings=n
Date
Hi,

The existing power saving loadbalancer CONFIG_SCHED_MC attempts to run
the workload in the system on minimum number of CPU packages and tries
to keep rest of the CPU packages idle for longer duration. Thus
consolidating workloads to fewer packages help other packages to be in
idle state and save power. The current implementation is very
conservative and does not work effectively across different workloads.
Initial idea of tunable sched_mc_power_savings=n was proposed to
enable tuning of the power saving load balancer based on the system
configuration, workload characteristics and end user requirements.

The power savings and performance of the given workload in an under
utilised system can be controlled by setting values of 0, 1 or 2 to
/sys/devices/system/cpu/sched_mc_power_savings with 0 being highest
performance and least power savings and level 2 indicating maximum
power savings even at the cost of slight performance degradation.

Please refer to the following discussions and article for details.

[1]Making power policy just work
http://lwn.net/Articles/287924/

[2][RFC v1] Tunable sched_mc_power_savings=n
http://lwn.net/Articles/287882/

v2: http://lwn.net/Articles/297306/
v3: http://lkml.org/lkml/2008/11/10/260
v4: http://lkml.org/lkml/2008/11/21/47
v5: http://lkml.org/lkml/2008/12/11/178
v6: http://lwn.net/Articles/311830/

The following series of patch demonstrates the basic framework for
tunable sched_mc_power_savings.

This version of the patch incorporates comments and feedback received
on the previous post from Andrew Morton.

Changes form v6:
----------------
* Convert BALANCE_FOR_xx_POWER and related macros to inline functions
based on comments from Andrew and Ingo.
* Ran basic kernelbench test and did not see any performance variation
due to the changes.
Changes form v5:
---------------
* Fixed the sscanf bug and checking for (p->flags & PF_KTHREAD)
* Dropped the RFC prefix to indicate that the patch is ready for
testing and inclusion
* Patch series against 2.6.28-rc8 kernel
Changes from v4:
----------------
* Conditionally added SD_BALANCE_NEWIDLE flag to MC and CPU level
domain to help task consolidation when sched_mc > 0
Removing this flag significantly reduces the number load_balance
calls and considerably slows down task consolidation and in effect
power savings.
Ingo and Mike Galbraith removed this flag to improve performance for
select benchmarks. I have added the flags only when power savings
is requested and restore to full performance mode if sched_mc=0

* Fixed few whitespace issues
* Patch series against 2.6.28-rc8 kernel

Changes from v3:
----------------
* Fixed the locking code with double_lock_balance() in
active-balance-newidle.patch

* Moved sched_mc_preferred_wakeup_cpu to root_domain structure so that
each partitioned sched domain will get independent nominated cpu

* More comments in active-balance-newidle.patch

* Reverted sched MC level and CPU level fine tuning in v2.6.28-rc4 for
now. These affect consolidation since SD_BALANCE_NEWIDLE is
removed. I will rework the tuning in the next iteration to
selectively enable them at sched_mc=2

* Patch series on 2.6.28-rc6 kernel

Changes from v2:
----------------
* Fixed locking order issue in active-balance new-idle
* Moved the wakeup biasing code to wake_idle() function and preserve
wake_affine function. Previous version would break wake affine in
order to aggressively consolidate tasks
* Removed sched_mc_preferred_wakeup_cpu global variable and moved to
doms_cur/dattr_cur and added a per_cpu pointer to appropriate
storage in partitioned sched domain. This changed is needed to
preserve functionality in case of partitioned sched domains
* Patch on 2.6.28-rc3 kernel

Results:
--------
Basic functionality of the code has not changed and the power vs
performance benefits for kernbench are similar to the ones posted
earlier.

KERNBENCH Runs: make -j4 on a x86 8 core, dual socket quad core cpu
package system

SchedMC Run Time Package Idle Energy Power
0 81.68 52.83% 54.71% 1.00x J 1.00y W
1 80.70 36.62% 70.11% 0.95x J 0.96y W
2 74.95 19.53% 85.92% 0.90x J 0.98y W

The results are marginally better than the previous version of the
patch series which could be within the test variation.

Please feel free to test, and let me know your comments and feedback.
I will post more experimental results with various benchmarks.

Hi Ingo,

Please review and include in sched development tree for testing.

Thanks,
Vaidy

Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>

---
Gautham R Shenoy (1):
sched: Framework for sched_mc/smt_power_savings=N

Vaidyanathan Srinivasan (7):
sched: idle_balance() does not call load_balance_newidle()
sched: add SD_BALANCE_NEWIDLE at MC and CPU level for sched_mc>0
sched: activate active load balancing in new idle cpus
sched: bias task wakeups to preferred semi-idle packages
sched: nominate preferred wakeup cpu
sched: favour lower logical cpu number for sched_mc balance
sched: convert BALANCE_FOR_xx_POWER to inline functions


include/linux/sched.h | 57 +++++++++++++++++++++++++----
include/linux/topology.h | 6 ++-
kernel/sched.c | 89 +++++++++++++++++++++++++++++++++++++++++++---
kernel/sched_fair.c | 18 +++++++++
4 files changed, 153 insertions(+), 17 deletions(-)
--


\
 
 \ /
  Last update: 2008-12-18 18:55    [from the cache]
©2003-2011 Jasper Spaans