Sign in with
Sign up | Sign in
Your question

CPU Heat Throttling

Tags:
Last response: in Laptops & Notebooks
Share
Anonymous
a b D Laptop
September 10, 2004 6:10:37 AM

Archived from groups: comp.sys.laptops (More info?)

I believed that I have discovered something significant about small laptops.
In small format laptops like the Toshiba 3500 TabletPC, Toshiba advertises
it as a 1.3 GHz Pentium M. One knows/assumes that while running from
batteries the processor may be run at slower speeds to conserve power. That
is an expected and positive feature.

One also assumes that when running from AC power and the machine power
saving
modes are therefore in "full power" mode then one expects full 1.3 GHz
Pentium M
performance. Intel's Pentium M processor technology includes thermal sensing
such that a CPU will self throttle when it gets too hot. What I've found
with a Toshiba
3500 I've been working with is that the cooling provided in the basic design
is
insufficient to EVER run the CPU at full speed in a 25C room sitting on a
wood
desk for more than a half minute and usually less. I suspect that this
behavior
is endemic to all small format laptops. I'm simply claiming that advertising
these as 1.3 GHz Pentium M is consumer fraud as in the basic design they can
never run at best HALF that for any sustained period.

I noticed this while doing Acronis backups pushing over a LAN with the
compression set high. If I folded the 3500 at 90 degrees and set it on edge
and cooled the room then I could more than double the throughput but still
got only about half of what a PIII/1000 would do.

Does anyone know of a utility or way to in real time show what the CPU speed
really is and/or what the CPU temperature throttling situation is??

More about : cpu heat throttling

Anonymous
a b D Laptop
September 10, 2004 6:10:38 AM

Archived from groups: comp.sys.laptops (More info?)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Ron,

You might want to try MobileMeter to monitor your computers hardware.

Ciao . . . C.Joseph

That which a man buys too cheaply . . .
~ He esteems too lightly



Ron Reaugh wrote:

|I believed that I have discovered something significant about small
laptops.
|In small format laptops like the Toshiba 3500 TabletPC, Toshiba
advertises
|it as a 1.3 GHz Pentium M. One knows/assumes that while running from
|batteries the processor may be run at slower speeds to conserve
power. That
|is an expected and positive feature.
|
|One also assumes that when running from AC power and the machine power
|saving
|modes are therefore in "full power" mode then one expects full 1.3 GHz
|Pentium M
|performance. Intel's Pentium M processor technology includes thermal
sensing
|such that a CPU will self throttle when it gets too hot. What I've found
|with a Toshiba
|3500 I've been working with is that the cooling provided in the basic
design
|is
|insufficient to EVER run the CPU at full speed in a 25C room sitting on a
|wood
|desk for more than a half minute and usually less. I suspect that this
|behavior
|is endemic to all small format laptops. I'm simply claiming that
advertising
|these as 1.3 GHz Pentium M is consumer fraud as in the basic design
they can
|never run at best HALF that for any sustained period.
|
|I noticed this while doing Acronis backups pushing over a LAN with the
|compression set high. If I folded the 3500 at 90 degrees and set it
on edge
|and cooled the room then I could more than double the throughput but
still
|got only about half of what a PIII/1000 would do.
|
|Does anyone know of a utility or way to in real time show what the
CPU speed
|really is and/or what the CPU temperature throttling situation is??
|
|
|
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (MingW32)

iD8DBQFBQQ+B6bFq6mlbLOwRAjkXAJ9185xKIufxuTUB+NRULH2cVBICjQCgtIhs
Cdb8Y2l9P8KXWP2tXUgw+64=
=xFce
-----END PGP SIGNATURE-----
Anonymous
a b D Laptop
September 10, 2004 6:12:34 AM

Archived from groups: comp.sys.laptops (More info?)

Ron Reaugh wrote:

> One also assumes that when running from AC power and the machine power
> saving
> modes are therefore in "full power" mode then one expects full 1.3 GHz
> Pentium M
> performance. Intel's Pentium M processor technology includes thermal sensing
> such that a CPU will self throttle when it gets too hot. What I've found
> with a Toshiba
> 3500 I've been working with is that the cooling provided in the basic design
> is
> insufficient to EVER run the CPU at full speed in a 25C room sitting on a
> wood
> desk for more than a half minute and usually less. I suspect that this
> behavior
> is endemic to all small format laptops. I'm simply claiming that advertising
> these as 1.3 GHz Pentium M is consumer fraud as in the basic design they can
> never run at best HALF that for any sustained period.

> Does anyone know of a utility or way to in real time show what the CPU speed
> really is and/or what the CPU temperature throttling situation is??

Try SpeedswitchXP, it shows CPU speed and CPU load in realtime. Also, it
gives you direct low-level control over the throttling behavior,
although I haven't seen much documentation on this - you need to read
the specs on ACPI to really have the background for this.

If you're really determined, I think you can use WPCREDIT to edit the
power management registers on the 855 chipset. I've written a .PCR file
for the 855 PCI Device #0 (Memory controller). The power management
registers are on PCI Device #6, I haven't gone there yet but it's all
documented in the Intel 82855 data sheets, you can get the PDF from
Intel's web site. It may prove to be a dead-end though; these registers
have a write-protect bit such that they can only be written once (by the
BIOS, at boot time) and then never changed until the next reset.
--
-- Howard Chu
Chief Architect, Symas Corp. Director, Highland Sun
http://www.symas.com http://highlandsun.com/hyc
Symas: Premier OpenSource Development and Support
Related resources
September 10, 2004 9:33:07 AM

Archived from groups: comp.sys.laptops (More info?)

Ron Reaugh <rondashreaugh@att.net> wrote:
: I believed that I have discovered something significant about small laptops.
: In small format laptops like the Toshiba 3500 TabletPC, Toshiba advertises
: it as a 1.3 GHz Pentium M. One knows/assumes that while running from
: batteries the processor may be run at slower speeds to conserve power. That
: is an expected and positive feature.

: One also assumes that when running from AC power and the machine power
: saving modes are therefore in "full power" mode then one expects
: full 1.3 GHz Pentium M performance.

Actually, Toshiba laptops have a utility to allow you to set power
management scehemes (CPU speed, etc.) for AC Power, Battery, etc. You
can set your laptop to run at "Long Life" when on Battery or full
speed if you really want maximum CPU power.

Go into Control Panel and look for "Toshiba Power Saver" to see what
your power settings are set to.

: Intel's Pentium M processor technology includes thermal sensing
: such that a CPU will self throttle when it gets too hot.

Actually, even Pentium 4 CPU's include throttling. My Desktop Pentium
4 Prescott system allows me to set the throttling temperature in the
BIOS. I have no idea if you have this flexibility on your laptop, but
you could try getting into the BIOS to check.

: What I've found
: with a Toshiba 3500 I've been working with is that the cooling
: provided in the basic design is insufficient to EVER run the CPU at
: full speed in a 25C room sitting on a wood desk for more than a
: half minute and usually less. I suspect that this behavior is
: endemic to all small format laptops. I'm simply claiming that
: advertising these as 1.3 GHz Pentium M is consumer fraud as in the
: basic design they can never run at best HALF that for any sustained
: period.

Sounds unlikely. Pentium M should be a pretty cool CPU at 1.3GHZ.

: I noticed this while doing Acronis backups pushing over a LAN with the
: compression set high. If I folded the 3500 at 90 degrees and set it on edge
: and cooled the room then I could more than double the throughput but still
: got only about half of what a PIII/1000 would do.

I'm not sure what you mean. You think the CPU is throttling down
because throughput is bad? Or that your througput improves when you
cool the room? How did you conduct such tests?

: Does anyone know of a utility or way to in real time show what the CPU speed
: really is and/or what the CPU temperature throttling situation is??

Not sure. But I do have something installed on my desktop computer (I
think it's "Motherboard Monitor") to monitor the CPU temperature, and
since I can set the throttle temp in the BIOS, I can see when it hits
the throttle temp (Prescott runs VERY HOT!!!). Running a
CPU-intensive task, I can notice the slowdown when it hits throttle
temp so I can see it is happening.

Andrew
--
----> Portland, Oregon, USA <----
*******************************************************************
----> http://www.bizave.com <---- Photo Albums and Portland Info
----> To Email me remove "MYSHOES" from email address
*******************************************************************
Anonymous
a b D Laptop
September 10, 2004 10:31:23 AM

Archived from groups: comp.sys.laptops (More info?)

"Andrew" <usenetMYSHOES@bizaveMYSHOES.com> wrote in message
news:zcasdaxchjzcasdac66124450609@bizaveMYSHOES.com...
> Ron Reaugh <rondashreaugh@att.net> wrote:
> : I believed that I have discovered something significant about small
laptops.
> : In small format laptops like the Toshiba 3500 TabletPC, Toshiba
advertises
> : it as a 1.3 GHz Pentium M. One knows/assumes that while running from
> : batteries the processor may be run at slower speeds to conserve power.
That
> : is an expected and positive feature.
>
> : One also assumes that when running from AC power and the machine power
> : saving modes are therefore in "full power" mode then one expects
> : full 1.3 GHz Pentium M performance.
>
> Actually, Toshiba laptops have a utility to allow you to set power
> management scehemes (CPU speed, etc.) for AC Power, Battery, etc.

Of course. I already described that.

> You
> can set your laptop to run at "Long Life" when on Battery or full
> speed if you really want maximum CPU power.

Yes, you can make such a setting. The point is that that setting does NOT
do what one might think it does.

> Go into Control Panel and look for "Toshiba Power Saver" to see what
> your power settings are set to.

You need to go back and read my original post carefully.

> : Intel's Pentium M processor technology includes thermal sensing
> : such that a CPU will self throttle when it gets too hot.
>
> Actually, even Pentium 4 CPU's include throttling. My Desktop Pentium
> 4 Prescott system allows me to set the throttling temperature in the
> BIOS. I have no idea if you have this flexibility on your laptop, but
> you could try getting into the BIOS to check.

The point is that there is a Intel HW set upper temperature limit at which
point the CPU starts throttling regardless.

> : What I've found
> : with a Toshiba 3500 I've been working with is that the cooling
> : provided in the basic design is insufficient to EVER run the CPU at
> : full speed in a 25C room sitting on a wood desk for more than a
> : half minute and usually less. I suspect that this behavior is
> : endemic to all small format laptops. I'm simply claiming that
> : advertising these as 1.3 GHz Pentium M is consumer fraud as in the
> : basic design they can never run at best HALF that for any sustained
> : period.
>
> Sounds unlikely. Pentium M should be a pretty cool CPU at 1.3GHZ.

Nevertheless those are the results and such is apparently confirmed by
others but not yet fully understood..

> : I noticed this while doing Acronis backups pushing over a LAN with the
> : compression set high. If I folded the 3500 at 90 degrees and set it on
edge
> : and cooled the room then I could more than double the throughput but
still
> : got only about half of what a PIII/1000 would do.
>
> I'm not sure what you mean. You think the CPU is throttling down
> because throughput is bad?

I'm certain of it. At max compression(100% CPU intensive) the backup rate
is directly proportional to cooling.

> Or that your througput improves when you
> cool the room? How did you conduct such tests?

That's described in my original post. As soon as I discovered this I was
able to remeber how this temperature/cooling throttle has been affecting
operation all along. Further there is a long outstanding issue with these
Tablets of jerky pausing operation that no one has ever identified until I
discovered this. I suspect similar affects many small format/compact
laptops where sufficient cooling can't be provided for full continuous CPU
max speed.

> : Does anyone know of a utility or way to in real time show what the CPU
speed
> : really is and/or what the CPU temperature throttling situation is??
>
> Not sure. But I do have something installed on my desktop computer (I
> think it's "Motherboard Monitor") to monitor the CPU temperature, and
> since I can set the throttle temp in the BIOS, I can see when it hits
> the throttle temp (Prescott runs VERY HOT!!!). Running a
> CPU-intensive task, I can notice the slowdown when it hits throttle
> temp so I can see it is happening.

Get a bigger heatsink and fan.
Anonymous
a b D Laptop
September 10, 2004 1:08:44 PM

Archived from groups: comp.sys.laptops (More info?)

Ron Reaugh wrote:
> "Andrew" <usenetMYSHOES@bizaveMYSHOES.com> wrote in message
> news:zcasdaxchjzcasdac66124450609@bizaveMYSHOES.com...
>> Ron Reaugh <rondashreaugh@att.net> wrote:
>>> I believed that I have discovered something significant about small
>>> laptops. In small format laptops like the Toshiba 3500 TabletPC,
>>> Toshiba advertises it as a 1.3 GHz Pentium M. One knows/assumes
>>> that while running from batteries the processor may be run at
>>> slower speeds to conserve power. That is an expected and positive
>>> feature.
>>
>>> One also assumes that when running from AC power and the machine
>>> power saving modes are therefore in "full power" mode then one
>>> expects
>>> full 1.3 GHz Pentium M performance.
>>
>> Actually, Toshiba laptops have a utility to allow you to set power
>> management scehemes (CPU speed, etc.) for AC Power, Battery, etc.
>
> Of course. I already described that.
>
>> You
>> can set your laptop to run at "Long Life" when on Battery or full
>> speed if you really want maximum CPU power.
>
> Yes, you can make such a setting. The point is that that setting
> does NOT do what one might think it does.
>
>> Go into Control Panel and look for "Toshiba Power Saver" to see what
>> your power settings are set to.
>
> You need to go back and read my original post carefully.
>
>>> Intel's Pentium M processor technology includes thermal sensing
>>> such that a CPU will self throttle when it gets too hot.
>>
>> Actually, even Pentium 4 CPU's include throttling. My Desktop
>> Pentium 4 Prescott system allows me to set the throttling
>> temperature in the BIOS. I have no idea if you have this
>> flexibility on your laptop, but you could try getting into the BIOS
>> to check.
>
> The point is that there is a Intel HW set upper temperature limit at
> which point the CPU starts throttling regardless.
>
>>> What I've found
>>> with a Toshiba 3500 I've been working with is that the cooling
>>> provided in the basic design is insufficient to EVER run the CPU at
>>> full speed in a 25C room sitting on a wood desk for more than a
>>> half minute and usually less. I suspect that this behavior is
>>> endemic to all small format laptops. I'm simply claiming that
>>> advertising these as 1.3 GHz Pentium M is consumer fraud as in the
>>> basic design they can never run at best HALF that for any sustained
>>> period.
>>
>> Sounds unlikely. Pentium M should be a pretty cool CPU at 1.3GHZ.
>
> Nevertheless those are the results and such is apparently confirmed by
> others but not yet fully understood..
>
>>> I noticed this while doing Acronis backups pushing over a LAN with
>>> the compression set high. If I folded the 3500 at 90 degrees and
>>> set it on edge and cooled the room then I could more than double
>>> the throughput but still got only about half of what a PIII/1000
>>> would do.
>>
>> I'm not sure what you mean. You think the CPU is throttling down
>> because throughput is bad?
>
> I'm certain of it. At max compression(100% CPU intensive) the backup
> rate is directly proportional to cooling.
>
>> Or that your througput improves when you
>> cool the room? How did you conduct such tests?
>
> That's described in my original post. As soon as I discovered this I
> was able to remeber how this temperature/cooling throttle has been
> affecting operation all along. Further there is a long outstanding
> issue with these Tablets of jerky pausing operation that no one has
> ever identified until I discovered this. I suspect similar affects
> many small format/compact laptops where sufficient cooling can't be
> provided for full continuous CPU max speed.
>
>>> Does anyone know of a utility or way to in real time show what the
>>> CPU speed really is and/or what the CPU temperature throttling
>>> situation is??
>>
>> Not sure. But I do have something installed on my desktop computer
>> (I think it's "Motherboard Monitor") to monitor the CPU temperature,
>> and since I can set the throttle temp in the BIOS, I can see when it
>> hits the throttle temp (Prescott runs VERY HOT!!!). Running a
>> CPU-intensive task, I can notice the slowdown when it hits throttle
>> temp so I can see it is happening.
>
> Get a bigger heatsink and fan.

It is not clear from your OP if when you use the Toshiba power panel
rather than Windows power settings, that you leave the Toshiba applet
running. Some of the OEM power applets, actually supplied by the BIOS
vendor, must be running (in the system tray) to be effective.

Q
Anonymous
a b D Laptop
September 10, 2004 2:33:48 PM

Archived from groups: comp.sys.laptops (More info?)

I can confirm such a behavior with my fanless Sony X505 (1.1 GHz Pentium M
ULV). It is not able to run at full speed for a sustained period of time
either. The problem is that the cooling system of such machines is not
designed for dissipating the heat from the processor at full speed.

As to consumer fraud, at least Sony US has a footnote in the specs saying
that the 1.1 GHz processor speed will not be available under all operating
conditions.

It seems that notebook reviews should take more care of this issue. If they
only gauge the fan noise, we might see more notebooks that use throtteling
instead of proper cooling in the future.

> Does anyone know of a utility or way to in real time show what the CPU
speed
> really is and/or what the CPU temperature throttling situation is??

You may try SpeedstepXP, which also lets you tweak many "hidden" power
management options.


"Ron Reaugh" <rondashreaugh@att.net> schrieb im Newsbeitrag
news:x280d.338290$OB3.139501@bgtnsc05-news.ops.worldnet.att.net...
> I believed that I have discovered something significant about small
laptops.
> In small format laptops like the Toshiba 3500 TabletPC, Toshiba advertises
> it as a 1.3 GHz Pentium M. One knows/assumes that while running from
> batteries the processor may be run at slower speeds to conserve power.
That
> is an expected and positive feature.
>
> One also assumes that when running from AC power and the machine power
> saving
> modes are therefore in "full power" mode then one expects full 1.3 GHz
> Pentium M
> performance. Intel's Pentium M processor technology includes thermal
sensing
> such that a CPU will self throttle when it gets too hot. What I've found
> with a Toshiba
> 3500 I've been working with is that the cooling provided in the basic
design
> is
> insufficient to EVER run the CPU at full speed in a 25C room sitting on a
> wood
> desk for more than a half minute and usually less. I suspect that this
> behavior
> is endemic to all small format laptops. I'm simply claiming that
advertising
> these as 1.3 GHz Pentium M is consumer fraud as in the basic design they
can
> never run at best HALF that for any sustained period.
>
> I noticed this while doing Acronis backups pushing over a LAN with the
> compression set high. If I folded the 3500 at 90 degrees and set it on
edge
> and cooled the room then I could more than double the throughput but still
> got only about half of what a PIII/1000 would do.
>
> Does anyone know of a utility or way to in real time show what the CPU
speed
> really is and/or what the CPU temperature throttling situation is??
>
>
>
Anonymous
a b D Laptop
September 10, 2004 10:33:04 PM

Archived from groups: comp.sys.laptops (More info?)

"Crazier" <crazier35@hotmail.com> wrote in message
news:414166ec$0$416$9b4e6d93@newsread2.arcor-online.net...
> I can confirm such a behavior with my fanless Sony X505 (1.1 GHz Pentium M
> ULV). It is not able to run at full speed for a sustained period of time
> either.

How did you come to realize this?

The problem is that the cooling system of such machines is not
> designed for dissipating the heat from the processor at full speed.

EXACTLY. Without major caveats in the sales literature that's consumer
fraud.

> As to consumer fraud, at least Sony US has a footnote in the specs saying
> that the 1.1 GHz processor speed will not be available under all operating
> conditions.

Still grossly misleading as it should read "will not be available under any
sustained normal operating conditions."

> It seems that notebook reviews should take more care of this issue. If
they
> only gauge the fan noise, we might see more notebooks that use throtteling
> instead of proper cooling in the future.

I still believe it's fraud without a clearer statement of the limitation.

> > Does anyone know of a utility or way to in real time show what the CPU
> speed
> > really is and/or what the CPU temperature throttling situation is??
>
> You may try SpeedstepXP, which also lets you tweak many "hidden" power
> management options.
>
>
> "Ron Reaugh" <rondashreaugh@att.net> schrieb im Newsbeitrag
> news:x280d.338290$OB3.139501@bgtnsc05-news.ops.worldnet.att.net...
> > I believed that I have discovered something significant about small
> laptops.
> > In small format laptops like the Toshiba 3500 TabletPC, Toshiba
advertises
> > it as a 1.3 GHz Pentium M. One knows/assumes that while running from
> > batteries the processor may be run at slower speeds to conserve power.
> That
> > is an expected and positive feature.
> >
> > One also assumes that when running from AC power and the machine power
> > saving
> > modes are therefore in "full power" mode then one expects full 1.3 GHz
> > Pentium M
> > performance. Intel's Pentium M processor technology includes thermal
> sensing
> > such that a CPU will self throttle when it gets too hot. What I've found
> > with a Toshiba
> > 3500 I've been working with is that the cooling provided in the basic
> design
> > is
> > insufficient to EVER run the CPU at full speed in a 25C room sitting on
a
> > wood
> > desk for more than a half minute and usually less. I suspect that this
> > behavior
> > is endemic to all small format laptops. I'm simply claiming that
> advertising
> > these as 1.3 GHz Pentium M is consumer fraud as in the basic design they
> can
> > never run at best HALF that for any sustained period.
> >
> > I noticed this while doing Acronis backups pushing over a LAN with the
> > compression set high. If I folded the 3500 at 90 degrees and set it on
> edge
> > and cooled the room then I could more than double the throughput but
still
> > got only about half of what a PIII/1000 would do.
> >
> > Does anyone know of a utility or way to in real time show what the CPU
> speed
> > really is and/or what the CPU temperature throttling situation is??
> >
> >
> >
>
>
Anonymous
a b D Laptop
September 10, 2004 10:40:46 PM

Archived from groups: comp.sys.laptops (More info?)

"Quaoar" <quaoar@tenthplanet.net> wrote in message
news:VKWdnYfrZssPXtzcRVn-gg@comcast.com...
> Ron Reaugh wrote:
> > "Andrew" <usenetMYSHOES@bizaveMYSHOES.com> wrote in message
> > news:zcasdaxchjzcasdac66124450609@bizaveMYSHOES.com...
> >> Ron Reaugh <rondashreaugh@att.net> wrote:
> >>> I believed that I have discovered something significant about small
> >>> laptops. In small format laptops like the Toshiba 3500 TabletPC,
> >>> Toshiba advertises it as a 1.3 GHz Pentium M. One knows/assumes
> >>> that while running from batteries the processor may be run at
> >>> slower speeds to conserve power. That is an expected and positive
> >>> feature.
> >>
> >>> One also assumes that when running from AC power and the machine
> >>> power saving modes are therefore in "full power" mode then one
> >>> expects
> >>> full 1.3 GHz Pentium M performance.
> >>
> >> Actually, Toshiba laptops have a utility to allow you to set power
> >> management scehemes (CPU speed, etc.) for AC Power, Battery, etc.
> >
> > Of course. I already described that.
> >
> >> You
> >> can set your laptop to run at "Long Life" when on Battery or full
> >> speed if you really want maximum CPU power.
> >
> > Yes, you can make such a setting. The point is that that setting
> > does NOT do what one might think it does.
> >
> >> Go into Control Panel and look for "Toshiba Power Saver" to see what
> >> your power settings are set to.
> >
> > You need to go back and read my original post carefully.
> >
> >>> Intel's Pentium M processor technology includes thermal sensing
> >>> such that a CPU will self throttle when it gets too hot.
> >>
> >> Actually, even Pentium 4 CPU's include throttling. My Desktop
> >> Pentium 4 Prescott system allows me to set the throttling
> >> temperature in the BIOS. I have no idea if you have this
> >> flexibility on your laptop, but you could try getting into the BIOS
> >> to check.
> >
> > The point is that there is a Intel HW set upper temperature limit at
> > which point the CPU starts throttling regardless.
> >
> >>> What I've found
> >>> with a Toshiba 3500 I've been working with is that the cooling
> >>> provided in the basic design is insufficient to EVER run the CPU at
> >>> full speed in a 25C room sitting on a wood desk for more than a
> >>> half minute and usually less. I suspect that this behavior is
> >>> endemic to all small format laptops. I'm simply claiming that
> >>> advertising these as 1.3 GHz Pentium M is consumer fraud as in the
> >>> basic design they can never run at best HALF that for any sustained
> >>> period.
> >>
> >> Sounds unlikely. Pentium M should be a pretty cool CPU at 1.3GHZ.
> >
> > Nevertheless those are the results and such is apparently confirmed by
> > others but not yet fully understood..
> >
> >>> I noticed this while doing Acronis backups pushing over a LAN with
> >>> the compression set high. If I folded the 3500 at 90 degrees and
> >>> set it on edge and cooled the room then I could more than double
> >>> the throughput but still got only about half of what a PIII/1000
> >>> would do.
> >>
> >> I'm not sure what you mean. You think the CPU is throttling down
> >> because throughput is bad?
> >
> > I'm certain of it. At max compression(100% CPU intensive) the backup
> > rate is directly proportional to cooling.
> >
> >> Or that your througput improves when you
> >> cool the room? How did you conduct such tests?
> >
> > That's described in my original post. As soon as I discovered this I
> > was able to remeber how this temperature/cooling throttle has been
> > affecting operation all along. Further there is a long outstanding
> > issue with these Tablets of jerky pausing operation that no one has
> > ever identified until I discovered this. I suspect similar affects
> > many small format/compact laptops where sufficient cooling can't be
> > provided for full continuous CPU max speed.
> >
> >>> Does anyone know of a utility or way to in real time show what the
> >>> CPU speed really is and/or what the CPU temperature throttling
> >>> situation is??
> >>
> >> Not sure. But I do have something installed on my desktop computer
> >> (I think it's "Motherboard Monitor") to monitor the CPU temperature,
> >> and since I can set the throttle temp in the BIOS, I can see when it
> >> hits the throttle temp (Prescott runs VERY HOT!!!). Running a
> >> CPU-intensive task, I can notice the slowdown when it hits throttle
> >> temp so I can see it is happening.
> >
> > Get a bigger heatsink and fan.
>
> It is not clear from your OP if when you use the Toshiba power panel
> rather than Windows power settings, that you leave the Toshiba applet
> running. Some of the OEM power applets, actually supplied by the BIOS
> vendor, must be running (in the system tray) to be effective.

All standard Toshiba stuff is running. This issue has nothing to do with
power settings. Note the new post in this thread regarding the Sony X505.
In my opinion this issue borders on consumer fraud. CPUs on these machines
will NOT run continuously at full speed under normal operating conditions.
One must rethink how such machines will get used when connected to AC power.
They are not able to be used like a 1.3GHz PC but more like a PIII/500 for
continous CPU saturation.
Anonymous
a b D Laptop
September 11, 2004 5:28:11 AM

Archived from groups: comp.sys.laptops (More info?)

> > I can confirm such a behavior with my fanless Sony X505 (1.1 GHz Pentium
M
> > ULV). It is not able to run at full speed for a sustained period of time
> > either.

> How did you come to realize this?

I saw a note in the X505's manual advising not to run processor-intense
programs such as some screensavers, as this may shorten the CPU's lifetime.
This was a surprise for me, as I thought computers were made for processing.
I presumed that there might be a heat/cooling issue, ran some benchmarks
while measuring CPU clock and temperature. This revealed the problem.

> > The problem is that the cooling system of such machines is not
> > designed for dissipating the heat from the processor at full speed.

> EXACTLY. Without major caveats in the sales literature that's consumer
> fraud.

I must say that I felt this way too after realizing that the machine would
not live up to its specification. In germany there was no footnote in the
specs. Then I wrote a program to find out the average and minimum
performance. On a hot day, the average performance turned out to be around
90%, minimum was about 70%, with 1.1GHz = 100%. I decided for me that a 10%
average drop would be acceptable and that a 70% minimum might still be
acceptable with realtime applications, given that they need some headroom
anyway. This is the way I calmed down myself, but it might not work for you
;-)

> > As to consumer fraud, at least Sony US has a footnote in the specs
saying
> > that the 1.1 GHz processor speed will not be available under all
operating
> > conditions.

> Still grossly misleading as it should read "will not be available under
any
> sustained normal operating conditions."

I'd suggest "The machine is not designed to run the CPU at its rated speed".

> I still believe it's fraud without a clearer statement of the limitation.

CPU speed is known to be an important feature for many users. Thus not
making very clear that a machine is not designed for the rated CPU speed is
at least a violation of acting in good faith.
Anonymous
a b D Laptop
September 11, 2004 5:28:12 AM

Archived from groups: comp.sys.laptops (More info?)

"Crazier" <crazier35@hotmail.com> wrote in message
news:4142388c$0$28478$9b4e6d93@newsread4.arcor-online.net...
> > > I can confirm such a behavior with my fanless Sony X505 (1.1 GHz
Pentium
> M
> > > ULV). It is not able to run at full speed for a sustained period of
time
> > > either.
>
> > How did you come to realize this?
>
> I saw a note in the X505's manual advising not to run processor-intense
> programs such as some screensavers, as this may shorten the CPU's
lifetime.
> This was a surprise for me, as I thought computers were made for
processing.
> I presumed that there might be a heat/cooling issue, ran some benchmarks
> while measuring CPU clock and temperature. This revealed the problem.
>
> > > The problem is that the cooling system of such machines is not
> > > designed for dissipating the heat from the processor at full speed.
>
> > EXACTLY. Without major caveats in the sales literature that's consumer
> > fraud.
>
> I must say that I felt this way too after realizing that the machine would
> not live up to its specification. In germany there was no footnote in the
> specs. Then I wrote a program to find out the average and minimum
> performance. On a hot day, the average performance turned out to be around
> 90%, minimum was about 70%, with 1.1GHz = 100%. I decided for me that a
10%

I find that in an average room 23C that this Toshiba 3500 is doing at best
50%.

> average drop would be acceptable and that a 70% minimum might still be
> acceptable with realtime applications, given that they need some headroom
> anyway. This is the way I calmed down myself, but it might not work for
you
> ;-)
>
> > > As to consumer fraud, at least Sony US has a footnote in the specs
> saying
> > > that the 1.1 GHz processor speed will not be available under all
> operating
> > > conditions.
>
> > Still grossly misleading as it should read "will not be available under
> any
> > sustained normal operating conditions."
>
> I'd suggest "The machine is not designed to run the CPU at its rated
speed".
>
> > I still believe it's fraud without a clearer statement of the
limitation.
>
> CPU speed is known to be an important feature for many users. Thus not
> making very clear that a machine is not designed for the rated CPU speed
is
> at least a violation of acting in good faith.

I'm in complete agreement. I have now gone back and carefully inpsected the
specs for both the Toshiba 3500 and newer M200 and BOTH have disclaimers
about CPU performance. However I NEVER would have read them to suggest
these findings. Only now after the discovery is it apparent. I also note
that HP/Compaq, Motion, Nec and Fujitsu Tablets do NOT have any such
footnotes.

Selling a docking station for a laptop with this deficiency is a further
fraud.
!