* [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off
@ 2013-11-02 21:42 bugzilla-daemon
2013-11-04 16:35 ` [Bug 64251] " bugzilla-daemon
` (5 more replies)
0 siblings, 6 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-02 21:42 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
Bug ID: 64251
Summary: Intel Pstate driver minimum percent does not change
with turbo off
Product: Power Management
Version: 2.5
Kernel Version: 3.12rc7
Hardware: All
OS: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: cpufreq
Assignee: cpufreq@vger.kernel.org
Reporter: dsmythies@telus.net
Regression: No
Created attachment 113141
--> https://bugzilla.kernel.org/attachment.cgi?id=113141&action=edit
Frequency Vs. Requested Frequency
When turbo is turned off the maximum frequency percent does not change. O.K.
So the minimum frequency percent should change, but it doesn't.
Using my computer as an example: The turbo on maximum frequency is 3.8 GHz; The
turbo off maximum frequency is 3.4 GHz; The minimum frequency is 1.6 GHz.
Therefore we would expect: For turbo on, max percent = 100, min percent = 42;
For turbo off, max percent = 100, min percent = 47.
starting from a fresh boot:
doug@s15:~$ uname -a
Linux s15 3.12.0-rc7+ #48 SMP Fri Nov 1 21:09:51 PDT 2013 x86_64 x86_64 x86_64
GNU/Linux
doug@s15:~$ cat /sys/devices/system/cpu/intel_pstate/*
100
42
0
doug@s15:~$ cd temp
doug@s15:~/temp$ sudo ./set_cpu_turbo_off
[sudo] password for doug:
doug@s15:~/temp$ cat /sys/devices/system/cpu/intel_pstate/*
100
42
1
doug@s15:~/temp$ cat set_cpu_turbo_off
#! /bin/bash
echo "1" > /sys/devices/system/cpu/intel_pstate/no_turbo
The attached png file demonstrates. (the rounding part will be a separate bug
report)
Note: This bug is extracted from bug 59481 which mentioned several issues, the
main issue was solved and the bug closed. Now I am following up on the other
issues.
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
@ 2013-11-04 16:35 ` bugzilla-daemon
2013-11-05 7:31 ` bugzilla-daemon
` (4 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-04 16:35 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
Dirk Brandewie <dirk.brandewie@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |dirk.brandewie@gmail.com
--- Comment #1 from Dirk Brandewie <dirk.brandewie@gmail.com> ---
min_pct_perf is a user configurable floor value (not to go below) so it is
still valid in both cases.
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
2013-11-04 16:35 ` [Bug 64251] " bugzilla-daemon
@ 2013-11-05 7:31 ` bugzilla-daemon
2013-11-05 7:56 ` bugzilla-daemon
` (3 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-05 7:31 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
--- Comment #2 from Doug Smythies <dsmythies@telus.net> ---
But the system overrides any user value that had been previously set upon mode
change. It doesn't make any sense (to me at least) in turbo off that it be set
to 42, which is basically a meaningless number under these conditions.
Example:
doug@s15:~/temp$ echo "53" | sudo tee
/sys/devices/system/cpu/intel_pstate/min_perf_pct
53
doug@s15:~/temp$ cat /sys/devices/system/cpu/intel_pstate/*
53
53
1
doug@s15:~/temp$ sudo ./set_cpu_performance
doug@s15:~/temp$ cat /sys/devices/system/cpu/intel_pstate/*
100
100
0
doug@s15:~/temp$ sudo ./set_cpu_powersave
doug@s15:~/temp$ cat /sys/devices/system/cpu/intel_pstate/*
53
42
0
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
2013-11-04 16:35 ` [Bug 64251] " bugzilla-daemon
2013-11-05 7:31 ` bugzilla-daemon
@ 2013-11-05 7:56 ` bugzilla-daemon
2013-11-07 5:28 ` bugzilla-daemon
` (2 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-05 7:56 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
--- Comment #3 from Dirk Brandewie <dirk.brandewie@gmail.com> ---
I was trying to provide some useful information by setting min_perf_pct to a
value. I guess I could leave it at 0 since any value between 0 and the percent
of the cpu represented by the lowest P state are the same.
The reason for having the ability to disable turbo is for people that avoid the
possible jitter introduced into the system caused by thermal throttling.
If you want to limit the P state selection to something to something greater
than the minimum to further reduce jitter caused by P state transitions then
you would set the min/floor value to something significantly higher than the
minimums.
Regardless 42% is a valid floor value.
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
` (2 preceding siblings ...)
2013-11-05 7:56 ` bugzilla-daemon
@ 2013-11-07 5:28 ` bugzilla-daemon
2013-11-07 5:36 ` bugzilla-daemon
2014-04-18 3:01 ` bugzilla-daemon
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-07 5:28 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
--- Comment #4 from Doug Smythies <dsmythies@telus.net> ---
(In reply to Dirk Brandewie from comment #3)
> I was trying to provide some useful information by setting min_perf_pct to a
> value.
Which I totally agree with, and do find the information very useful.
I am just suggesting that it would be equally useful to provide that same
information for the other cases.
I see 4 scenarios, maybe there are more (using my computer as the example):
1.) BIOS allows Turbo mode and Turbo on enabled in driver:
min = 42 percent, which represents 1.6 GHz and is accurate
max = 100 percent, which represents 3.8 GHz and is accurate
2.) BIOS allows Turbo mode and Turbo is disabled in driver:
min = 42 percent, which doesn't represent anything.
max = 100 percent, which represents 3.4 GHz and is accurate.
3.) BIOS disabled Turbo mode and Turbo on enabled in driver:
min = 42 percent, which represents 1.6 GHz and is accurate.
max = 100 percent, which represents 3.8 GHz, but can not be achieved.
(a supporting graph will be attached in a minute)
4.) BIOS disabled Turbo mode and Turbo is disabled in driver:
This does not work at all on my computer. The CPU frequency just locks at
whatever it was when the driver Turbo disable transition command is issued.
(a supporting graph will be attached in a minute)
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
` (3 preceding siblings ...)
2013-11-07 5:28 ` bugzilla-daemon
@ 2013-11-07 5:36 ` bugzilla-daemon
2014-04-18 3:01 ` bugzilla-daemon
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2013-11-07 5:36 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
--- Comment #5 from Doug Smythies <dsmythies@telus.net> ---
Created attachment 113721
--> https://bugzilla.kernel.org/attachment.cgi?id=113721&action=edit
BIOS disabled Turbo mode CPU freq Vs. Requested.
This graph is in support of the previous comments.
The "rnd" and "nokick" in the legend are related to other bug reports, and can
be ignored in the context of this bug report. They are both with driver turbo
on, but BIOS turbo disabled.
"Toff" means driver turbo disabled, and multiple tests were done, each one with
a different frequency on CPU 7 at the time of driver transition from turbo
enabled to disabled, to demonstrate that the CPU frequency will lock and be
thereafter unchangeable.
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug 64251] Intel Pstate driver minimum percent does not change with turbo off
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
` (4 preceding siblings ...)
2013-11-07 5:36 ` bugzilla-daemon
@ 2014-04-18 3:01 ` bugzilla-daemon
5 siblings, 0 replies; 7+ messages in thread
From: bugzilla-daemon @ 2014-04-18 3:01 UTC (permalink / raw)
To: cpufreq
https://bugzilla.kernel.org/show_bug.cgi?id=64251
Lan Tianyu <tianyu.lan@intel.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |tianyu.lan@intel.com
Assignee|cpufreq@vger.kernel.org |dirk.brandewie@gmail.com
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2014-04-18 3:01 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-11-02 21:42 [Bug 64251] New: Intel Pstate driver minimum percent does not change with turbo off bugzilla-daemon
2013-11-04 16:35 ` [Bug 64251] " bugzilla-daemon
2013-11-05 7:31 ` bugzilla-daemon
2013-11-05 7:56 ` bugzilla-daemon
2013-11-07 5:28 ` bugzilla-daemon
2013-11-07 5:36 ` bugzilla-daemon
2014-04-18 3:01 ` bugzilla-daemon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).