How to configure cpudynd - Revision history

Gandro at 15:30, 10 December 2008

2008-12-10T15:30:59Z

Wyrfel at 22:34, 17 January 2006

2006-01-17T22:34:47Z

194.39.131.39 at 07:34, 11 November 2005

2005-11-11T07:34:48Z

194.39.131.39 at 07:33, 11 November 2005

2005-11-11T07:33:47Z

84.162.5.28 at 19:46, 10 November 2005

2005-11-10T19:46:28Z

84.162.5.28 at 19:41, 10 November 2005

2005-11-10T19:41:19Z

84.162.5.28 at 19:39, 10 November 2005

2005-11-10T19:39:58Z

Pebolle: Use stub template

2005-09-30T19:43:06Z

Use stub template

142.167.41.155 at 15:49, 25 June 2005

2005-06-25T15:49:01Z

New page

← Older revision		Revision as of 15:30, 10 December 2008
Line 2:		Line 2:

	If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd will work by changing the kernels frequency scaling governors: it will load the powersave governor by default and exchange it against the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (<tt>cpufreq_performance</tt> and <tt>cpufreq_powersave</tt>) are loaded properly.		If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd will work by changing the kernels frequency scaling governors: it will load the powersave governor by default and exchange it against the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (<tt>cpufreq_performance</tt> and <tt>cpufreq_powersave</tt>) are loaded properly.
		+
		+	{{NOTE\|There is a "bug" in cpudynd 1.0.1: The interval which is passed with the <tt>-i</tt> parameter is handled as deciseconds instead of seconds, which causes unnecessary wakeups of the cpu. Use the <tt>-ms</tt> parameter with the value in milliseconds as a workarouond for this. }}

	In {{Debian}}, all you have to do is:		In {{Debian}}, all you have to do is:

@@ Line 1: / Line 1: @@
 {{NOTE|See the [[How_to_make_use_of_Dynamic_Frequency_Scaling|Dynamic Frequency Scaling HOWTO]] on how to avoid using cpu frequency scaling daemons by using the kernel builtin ondemand governor.}}
-In Debian, all you have to do is:
+In {{Debian}}, all you have to do is:
 :{{cmdroot|apt-get install cpudyn}}
+In most cases it will work out of the box.
 This is the default Debian config, but works well out of the box:

@@ Line 7: / Line 7: @@
 And in most cases it will work out of the box.
-If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd will worksby exchanging the scaling governor: it will load the powersave governor by default and exchange it by the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (<tt>cpufreq_performance</tt> and <tt>cpufreq_powersave</tt>) are loaded properly.
+If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd will work by exchanging the scaling governor: it will load the powersave governor by default and exchange it by the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (<tt>cpufreq_performance</tt> and <tt>cpufreq_powersave</tt>) are loaded properly.
 This is the default Debian config, but works well out of the box:

@@ Line 7: / Line 7: @@
 And in most cases it will work out of the box.
-If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd works by exchanging the scaling governor: it will load the powersave governor by default and exchange it by the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (cpufreq_performance and cpufreq_powersave) are loaded properly.
+If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd will worksby exchanging the scaling governor: it will load the powersave governor by default and exchange it by the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors can be used by the kernel, i.e. their modules (<tt>cpufreq_performance</tt> and <tt>cpufreq_powersave</tt>) are loaded properly.
 This is the default Debian config, but works well out of the box:

@@ Line 3: / Line 3: @@
 In Debian, all you have to do is:
-:{{cmdroot|apt-get install cpudynd}}
+:{{cmdroot|apt-get install cpudyn}}
 And in most cases it will work out of the box.

← Older revision		Revision as of 19:39, 10 November 2005
Line 1:		Line 1:
−	{{~~Stub~~}}	+	{{NOTE\|See the [[How_to_make_use_of_Dynamic_Frequency_Scaling\|Dynamic Frequency Scaling HOWTO]] on how to avoid using cpu frequency scaling daemons by using the kernel builtin ondemand governor.}}
		+
		+
		+	In Debian, all you have to do is:
		+	:{{cmdroot\|apt-get install cpudynd}}
		+
		+	And in most cases it will work out of the box.
		+
		+	If you decide not to use the ACPI throttling functionality (as you certainly will on simple SpeedStep models like all T2x), cpudynd works by exchanging the scaling governor: it will load the powersave governor by default and exchange it by the performance governor if the CPU load exceeds a certain level. Therefore, you have to make sure that those two governors are either compiled into the kernel or that they are loaded as modules (cpufreq_performance and cpufreq_powersave).
		+
		+	This is the default Debian config, but works well out of the box:
		+
		+	# Cpudyn configure options
		+
		+	################################################
		+	# COMMON OPTIONS
		+	################################################
		+	#
		+	# Internal between idle ratio tests in 1/10 sec increments
		+	#
		+	INTERVAL=1
		+
		+	################################################
		+	# CPUFREQ OPTIONS
		+	################################################
		+	#
		+	# CPU idle work ratio to speed up
		+	#
		+	CPU_UP=0.5
		+
		+	#
		+	# CPU idle work ratio to speed down
		+	#
		+	CPU_DOWN=0.9
		+
		+	#
		+	# when using cputhrottling, what state to switch when speeding down
		+	#
		+	THROTTLING_LOW=7
		+
		+	################################################
		+	# DISK OPTIONS
		+	# (disabled by default)
		+	################################################
		+	#
		+	# Timeout to put the disk in standby mode if there was no
		+	# io during that period (in seconds)
		+	#
		+	#TIMEOUT=120
		+
		+	#
		+	# Specified disks to spindown (comma separated devices)
		+	#
		+	# DISKS=/dev/hda,/dev/hdb
		+
		+	################################################
		+	# EXTRA OPTIONS
		+	################################################
		+
		+	# Available options are
		+	# -acpi disables cpufreq and force the use of acpi
		+	# -asus enables asus_acpi in some laptops
		+	# -nice count also nice CPU usage as load as well
		+	# -minf value Set the minimum CPU frequency in a value between 0.0 and 1.0
		+	#
		+	# EXTRA_OPTIONS=""

← Older revision		Revision as of 19:43, 30 September 2005
Line 1:		Line 1:
−	..	+	{{Stub}}