Re: [Powertop] HIgh power concumption on network interface

Re: [Powertop] HIgh power concumption on network interface

[Wout Mertens]

[-- Attachment #1: Type: text/plain, Size: 1197 bytes --]

kingma at bluewin.ch wrote on Wed Feb 12 23:32:27 PST 2014:

> I observe a sustained high power consumption of 5- 10 W on the ehternet
adaptor e1000e on a lenovo thinkgpad t430s. Such I high power drain cannot
be correct can it? the total power consumption is around 25 W which is a
concern for battery life.

I'm noticing the same thing on my Dell Latitude E6540. It also has an e1000
and powertop reports it as being the highest consumer, at around 15W. When
I unplug the system, the battery reports a drain of about 20W while idle
which is pretty bad IMHO. With the 68Wh battery it translates to about 2.5
hours battery life.

Lowest I've seen for battery drain is 14W, with the screen at lowest
brightness, Google Chrome turned off, disk spun down. I'm using the
built-in Intel graphics and the system has 16GB of ram and a quad-core i7.

I never ran windows on this laptop so I don't know what to expect but
reviews are claiming 7+ hours on the 90Wh battery. That would mean an
average of 13W, while doing actual work.

So I'm really wondering: Is the e1000 really eating up 15W or is powertop
misreporting?

I already calibrated powertop, without a change.

Wout.

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 1437 bytes --]

[Powertop] HIgh power concumption on network interface

[kingma]

[-- Attachment #1: Type: text/plain, Size: 247 bytes --]

I observe a sustained high power consumption of 5- 10 W on the ehternet adaptor e1000e on a lenovo thinkgpad t430s. Such I high power drain cannot be correct can it? the total power consumption is around 25 W which is a concern for battery life.

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 249 bytes --]

Re: [Powertop] HIgh power concumption on network interface

[Wout Mertens]

[-- Attachment #1: Type: text/plain, Size: 743 bytes --]

On Thu, Feb 20, 2014 at 4:06 PM, Arjan van de Ven <arjan(a)linux.intel.com>wrote:

> On 2/19/2014 10:35 PM, Wout Mertens wrote:
>
>> Alright I'll give that a shot, thanks!
>>
>> As a side question though, is 20W a reasonable idle power consumption for
>> a high-end recent laptop?
>>
>
> that's quite on the high side
> Having said that, screen takes a lot of power at full brightness ;)
>

Well this was at the lowest brightness I could set it at. I just tested,
and with the laptop going into "light sleep" (screen off, disk off, idle
processes) it seems to use 12W.

Any tips for finding out what might cause the drain? Perhaps the 16GB RAM
is suspect? Or maybe just the CPU (it's an i7-4800MQ with a 47W TDP).

Wout.

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 1383 bytes --]

Re: [Powertop] HIgh power concumption on network interface

[Arjan van de Ven]

[-- Attachment #1: Type: text/plain, Size: 296 bytes --]

On 2/19/2014 10:35 PM, Wout Mertens wrote:
> Alright I'll give that a shot, thanks!
>
> As a side question though, is 20W a reasonable idle power consumption for a high-end recent laptop?

that's quite on the high side
Having said that, screen takes a lot of power at full brightness ;)

Re: [Powertop] HIgh power concumption on network interface

[Sree Harsha Totakura]

[-- Attachment #1: Type: text/plain, Size: 258 bytes --]

On 02/20/2014 07:35 AM, Wout Mertens wrote:
> As a side question though, is 20W a reasonable idle power consumption for a
> high-end recent laptop?

In most cases, NO.  I have a T440s, it takes upto 7W with medium
brightness setting when idle.

Sree

Re: [Powertop] HIgh power concumption on network interface

[Wout Mertens]

[-- Attachment #1: Type: text/plain, Size: 2405 bytes --]

Alright I'll give that a shot, thanks!

As a side question though, is 20W a reasonable idle power consumption for a
high-end recent laptop?

Cheers,

Wout.
On Feb 19, 2014 7:53 PM, "Kok, Auke-jan H" <auke-jan.h.kok(a)intel.com> wrote:

>
>
>
> On Wed, Feb 19, 2014 at 1:13 AM, Wout Mertens <wout.mertens(a)gmail.com>wrote:
>
>> kingma at bluewin.ch wrote on Wed Feb 12 23:32:27 PST 2014:
>>
>> > I observe a sustained high power consumption of 5- 10 W on the ehternet
>> adaptor e1000e on a lenovo thinkgpad t430s. Such I high power drain cannot
>> be correct can it? the total power consumption is around 25 W which is a
>> concern for battery life.
>>
>> I'm noticing the same thing on my Dell Latitude E6540. It also has an
>> e1000 and powertop reports it as being the highest consumer, at around 15W.
>> When I unplug the system, the battery reports a drain of about 20W while
>> idle which is pretty bad IMHO. With the 68Wh battery it translates to about
>> 2.5 hours battery life.
>>
>> Lowest I've seen for battery drain is 14W, with the screen at lowest
>> brightness, Google Chrome turned off, disk spun down. I'm using the
>> built-in Intel graphics and the system has 16GB of ram and a quad-core i7.
>>
>> I never ran windows on this laptop so I don't know what to expect but
>> reviews are claiming 7+ hours on the 90Wh battery. That would mean an
>> average of 13W, while doing actual work.
>>
>> So I'm really wondering: Is the e1000 really eating up 15W or is powertop
>> misreporting?
>>
>
>
> There have been plenty of reports that point to the calibration being
> incomplete (Let's just call it that).
>
> Obviously e1000 won't eat 15W of power - 15W could power a 10gig Fiber
> Optic transceiver (a high-powered laser, basically). Power consumption is
> more in the order of a few watts max, with idle power much, much lower.
>
> Calibration could fail if during the calibration phase powertop is unable
> to toggle power to components or enable/disable devices. As a fallback, you
> can run the calibration multiple times and vary the enabled devices in the
> system. You could for instance go into the BIOS and disable the ethernet
> device entirely, calibrate, reboot and enable the ethernet device, then
> calibrate again. The extra data points should improve the power estimates
> for devices.
>
> Cheers,
>
> Auke
>
>

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 3325 bytes --]

Re: [Powertop] HIgh power concumption on network interface

[Kok, Auke-jan H]

[-- Attachment #1: Type: text/plain, Size: 2083 bytes --]

On Wed, Feb 19, 2014 at 1:13 AM, Wout Mertens <wout.mertens(a)gmail.com>wrote:

> kingma at bluewin.ch wrote on Wed Feb 12 23:32:27 PST 2014:
>
> > I observe a sustained high power consumption of 5- 10 W on the ehternet
> adaptor e1000e on a lenovo thinkgpad t430s. Such I high power drain cannot
> be correct can it? the total power consumption is around 25 W which is a
> concern for battery life.
>
> I'm noticing the same thing on my Dell Latitude E6540. It also has an
> e1000 and powertop reports it as being the highest consumer, at around 15W.
> When I unplug the system, the battery reports a drain of about 20W while
> idle which is pretty bad IMHO. With the 68Wh battery it translates to about
> 2.5 hours battery life.
>
> Lowest I've seen for battery drain is 14W, with the screen at lowest
> brightness, Google Chrome turned off, disk spun down. I'm using the
> built-in Intel graphics and the system has 16GB of ram and a quad-core i7.
>
> I never ran windows on this laptop so I don't know what to expect but
> reviews are claiming 7+ hours on the 90Wh battery. That would mean an
> average of 13W, while doing actual work.
>
> So I'm really wondering: Is the e1000 really eating up 15W or is powertop
> misreporting?
>


There have been plenty of reports that point to the calibration being
incomplete (Let's just call it that).

Obviously e1000 won't eat 15W of power - 15W could power a 10gig Fiber
Optic transceiver (a high-powered laser, basically). Power consumption is
more in the order of a few watts max, with idle power much, much lower.

Calibration could fail if during the calibration phase powertop is unable
to toggle power to components or enable/disable devices. As a fallback, you
can run the calibration multiple times and vary the enabled devices in the
system. You could for instance go into the BIOS and disable the ethernet
device entirely, calibrate, reboot and enable the ethernet device, then
calibrate again. The extra data points should improve the power estimates
for devices.

Cheers,

Auke

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 2800 bytes --]