* Inconsistent processor throttling behaviour on Clevo M720R/Intel T8100
@ 2009-05-01 15:43 The Holy ettlz
2009-05-04 1:15 ` Zhang Rui
0 siblings, 1 reply; 3+ messages in thread
From: The Holy ettlz @ 2009-05-01 15:43 UTC (permalink / raw)
To: linux-acpi
[-- Attachment #1: Type: text/plain, Size: 2673 bytes --]
Hello,
I'm encountering a problem on my notebook (Clevo M720R) with an Intel
T8100 processor whereby Core 0 is always in the T3 throttling state
following a reboot. There is no thermal reason why this should be so,
and it is quite annoying having to force all computationally-intensive
tasks onto Core 1 (which is unthrottled). This is reflected in the
benchmarks: Core 0 takes around 21s to do what Core 1 does in 12s.
In the /proc tree,
$ cat /proc/acpi/processor/CPU0/throttling
state count: 8
active state: T3
state available: T3 to T7
states:
T0: 100%
T1: 88%
T2: 75%
*T3: 63%
T4: 50%
T5: 38%
T6: 25%
T7: 13%
$ cat /proc/acpi/processor/CPU1/throttling
state count: 8
active state: T0
state available: T0 to T7
states:
*T0: 100%
T1: 88%
T2: 75%
T3: 63%
T4: 50%
T5: 38%
T6: 25%
T7: 13%
If I boot with acpi=off, there is no problem and I get full performance
from both cores. Now, if I suspend and resume, the situation changes.
The active state is outside the reported range:
$ cat /proc/acpi/processor/CPU0/throttling
state count: 8
active state: T8
state available: T3 to T7
states:
T0: 100%
T1: 88%
T2: 75%
T3: 63%
T4: 50%
T5: 38%
T6: 25%
T7: 13%
$ cat /proc/acpi/processor/CPU1/throttling
state count: 8
active state: T8
state available: T0 to T7
states:
T0: 100%
T1: 88%
T2: 75%
T3: 63%
T4: 50%
T5: 38%
T6: 25%
T7: 13%
I get full performance from both cores in this case.
Any suggestions, and what additional info would be useful? In the
meantime, is there any way to disable throttling without taking out the
rest of ACPI and thermal so that I can get my notebook's full
performance from the outset?
Many thanks,
James.
--
The Holy ettlz TheHolyettlz@googlemail.com
PGP key ID: 03F94B5D
-----------------------------------------------------------------------
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Inconsistent processor throttling behaviour on Clevo M720R/Intel T8100
2009-05-01 15:43 Inconsistent processor throttling behaviour on Clevo M720R/Intel T8100 The Holy ettlz
@ 2009-05-04 1:15 ` Zhang Rui
2009-05-06 20:01 ` The Holy ettlz
0 siblings, 1 reply; 3+ messages in thread
From: Zhang Rui @ 2009-05-04 1:15 UTC (permalink / raw)
To: The Holy ettlz; +Cc: linux-acpi@vger.kernel.org
On Fri, 2009-05-01 at 23:43 +0800, The Holy ettlz wrote:
> Hello,
>
> I'm encountering a problem on my notebook (Clevo M720R) with an Intel
> T8100 processor whereby Core 0 is always in the T3 throttling state
> following a reboot. There is no thermal reason why this should be so,
> and it is quite annoying having to force all computationally-intensive
> tasks onto Core 1 (which is unthrottled). This is reflected in the
> benchmarks: Core 0 takes around 21s to do what Core 1 does in 12s.
>
> In the /proc tree,
>
>
> $ cat /proc/acpi/processor/CPU0/throttling
> state count: 8
> active state: T3
> state available: T3 to T7
> states:
> T0: 100%
> T1: 88%
> T2: 75%
> *T3: 63%
> T4: 50%
> T5: 38%
> T6: 25%
> T7: 13%
>
please attach the output of "grep . /proc/acpi/processor/CPU*/*".
please attach the acpidump output.
> $ cat /proc/acpi/processor/CPU1/throttling
> state count: 8
> active state: T0
> state available: T0 to T7
> states:
> *T0: 100%
> T1: 88%
> T2: 75%
> T3: 63%
> T4: 50%
> T5: 38%
> T6: 25%
> T7: 13%
>
>
> If I boot with acpi=off, there is no problem and I get full performance
> from both cores. Now, if I suspend and resume, the situation changes.
> The active state is outside the reported range:
>
hmmm, I'll attach a debug patch after the acpidump file is available. :)
>
> $ cat /proc/acpi/processor/CPU0/throttling
> state count: 8
> active state: T8
> state available: T3 to T7
> states:
> T0: 100%
> T1: 88%
> T2: 75%
> T3: 63%
> T4: 50%
> T5: 38%
> T6: 25%
> T7: 13%
>
> $ cat /proc/acpi/processor/CPU1/throttling
> state count: 8
> active state: T8
> state available: T0 to T7
> states:
> T0: 100%
> T1: 88%
> T2: 75%
> T3: 63%
> T4: 50%
> T5: 38%
> T6: 25%
> T7: 13%
>
>
> I get full performance from both cores in this case.
>
> Any suggestions, and what additional info would be useful? In the
> meantime, is there any way to disable throttling without taking out the
> rest of ACPI and thermal so that I can get my notebook's full
> performance from the outset?
>
the answer is "no" so far.
But I think it's a good idea.
patch will be attached later.
thanks,
rui
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Inconsistent processor throttling behaviour on Clevo M720R/Intel T8100
2009-05-04 1:15 ` Zhang Rui
@ 2009-05-06 20:01 ` The Holy ettlz
0 siblings, 0 replies; 3+ messages in thread
From: The Holy ettlz @ 2009-05-06 20:01 UTC (permalink / raw)
To: Zhang Rui; +Cc: linux-acpi@vger.kernel.org
[-- Attachment #1: Type: text/plain, Size: 562 bytes --]
On Mon, 2009-05-04 at 09:15 +0800, Zhang Rui wrote:
>
> please attach the output of "grep . /proc/acpi/processor/CPU*/*".
> please attach the acpidump output.
Thanks for your response. Rather than clog up the mailing list with the
attachments, I've opened Bug 13259 at
http://bugzilla.kernel.org/show_bug.cgi?id=13259
with the requested information.
Regards,
James.
--
The Holy ettlz TheHolyettlz@googlemail.com
PGP key ID: 03F94B5D
-----------------------------------------------------------------------
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2009-05-06 20:01 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-05-01 15:43 Inconsistent processor throttling behaviour on Clevo M720R/Intel T8100 The Holy ettlz
2009-05-04 1:15 ` Zhang Rui
2009-05-06 20:01 ` The Holy ettlz
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox