athlon xp-m in home system

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

I'd like to upgrade one of my systems (Chaintech 7NIL1/nForce 2 mboard),
with a lower power processor, and I've read that the Athlon XP-M
processors work fine as long as the motherboard supports the
Barton core. Is that true? Is it just "drop-in" or Will I need
to set the processor voltage/speed/multiplier manually?
Does the BIOS need to be updated specifically for the XP-M?

Also, there appear to be different versions of the mobile processors.
What are the main differences?

Thanks,
Doug

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

dhs wrote:
> I'd like to upgrade one of my systems (Chaintech 7NIL1/nForce 2 mboard),
> with a lower power processor, and I've read that the Athlon XP-M
> processors work fine as long as the motherboard supports the
> Barton core. Is that true? Is it just "drop-in" or Will I need
> to set the processor voltage/speed/multiplier manually?
> Does the BIOS need to be updated specifically for the XP-M?

From everything I've heard, yes, it's true, just drop in the XP-M into
a desktop board.

> Also, there appear to be different versions of the mobile processors.
> What are the main differences?

I've heard that there is a low-power version of it, and then there is an
ultralow power. I'm not sure of the details, somebody else is going to
have to fill you in on these details.

Yousuf Khan

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:

>I'd like to upgrade one of my systems (Chaintech 7NIL1/nForce 2 mboard),
>with a lower power processor, and I've read that the Athlon XP-M
>processors work fine as long as the motherboard supports the
>Barton core. Is that true?

Almost. Some BIOSes will puke if you put an AthlonXP-M in the system,
most will work fine.

> Is it just "drop-in" or Will I need
>to set the processor voltage/speed/multiplier manually?

Depends on the motherboard. Voltage *should* be no problem,
multiplier probably won't be a problem, but bus is about a 50/50 shot
as to whether or not it will be set properly.

>Does the BIOS need to be updated specifically for the XP-M?

It doesn't need to specifically support the AthlonXP-M, but as
mentioned above, some BIOSes break when they see something they don't
recognize. Fortunately such BIOSes are relatively few and far
between, especially if you have the latest version.

>Also, there appear to be different versions of the mobile processors.
>What are the main differences?

That depends on which versions you are comparing. Some of the XP-M
chips run at 266MT/s bus speeds, some run at 333MT/s. Some have 256KB
of L2 cache, some have 512KB. Some have been tested to ensure that
their maximum TDP is 45W, others only 25W (the latter being damn near
impossible to find). In short, there are lots of possibilities,
though most can be safely ignored for 99% of uses.

My recommendation would be to fire your motherboard model into Google
along with AthlonXP-M and see what comes up.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

In article <utct819spdfljngjpm78v4k7j8g1vabkkp@4ax.com>,
Tony Hill <hilla_nospam_20@yahoo.ca> wrote:
>On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:
>> Is it just "drop-in" or Will I need
>>to set the processor voltage/speed/multiplier manually?
>
>Depends on the motherboard. Voltage *should* be no problem,

The motherboard is a uATX board that only allows setting the
FSB - no CPU voltage or multiplier settings. This is probably
a FAQ, but how does the BIOS set the voltage and multiplier
on Athlon machines? Does it read the CPU model and lookup
the correct voltage/multiplier in its own table, or does it
read these values directly from the processor? The latter
approach would have been better, but judging from the BIOS
upgrades that list CPU support as enhancements, it seems
that AMD used the former approach.

>multiplier probably won't be a problem, but bus is about a 50/50 shot
>as to whether or not it will be set properly.
The FSB bus speed can also be set by on-board jumpers.

>My recommendation would be to fire your motherboard model into Google
>along with AthlonXP-M and see what comes up.

I had tried this, but uATX boards don't sell very well, so there
aren't a lot of posting on this model.

Thanks!

Doug

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 21 May 2005 10:58:18 -0500, dhs@cs.utexas.edu (dhs) wrote:

>In article <utct819spdfljngjpm78v4k7j8g1vabkkp@4ax.com>,
>Tony Hill <hilla_nospam_20@yahoo.ca> wrote:
>>On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:
>>> Is it just "drop-in" or Will I need
>>>to set the processor voltage/speed/multiplier manually?
>>
>>Depends on the motherboard. Voltage *should* be no problem,
>
>The motherboard is a uATX board that only allows setting the
>FSB - no CPU voltage or multiplier settings. This is probably
>a FAQ, but how does the BIOS set the voltage and multiplier
>on Athlon machines? Does it read the CPU model and lookup
>the correct voltage/multiplier in its own table, or does it
>read these values directly from the processor? The latter
>approach would have been better, but judging from the BIOS
>upgrades that list CPU support as enhancements, it seems
>that AMD used the former approach.

It does indeed read the CPU model and lookup the correct
voltage/multiplier according to a set of fuses on the processor.
These fuses are actually visible and are set by AMD in
post-production. You'll often hear them referred to as "bridges", and
you can see where some have been cut by a laser on top of the
processor packaging.

>>multiplier probably won't be a problem, but bus is about a 50/50 shot
>>as to whether or not it will be set properly.
>The FSB bus speed can also be set by on-board jumpers.

You should be fine in that case. One of the desirable characteristics
of these AthlonXP-M processors is that you can usually bump the bus
speed up a notch and overclock the processor with no other changes to
the system. This often results in a ~20% increase in both bus speed
and processor speed for about 10 seconds worth of work.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca

 

[Andy]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:

>
>I'd like to upgrade one of my systems (Chaintech 7NIL1/nForce 2 mboard),
>with a lower power processor, and I've read that the Athlon XP-M
>processors work fine as long as the motherboard supports the
>Barton core. Is that true? Is it just "drop-in" or Will I need
>to set the processor voltage/speed/multiplier manually?

The BIOS will not recognize the processor, so the voltage and
multiplier have to be manually set. I used to run a mobile XP 2400+ on
an MSI KM3M-V uATX motherboard. The default core voltage was set to
1.7 volts, and there was no way to change it. The default multiplier
was set to 6, resulting in CPU clock speed of 800MHz. Using the
program CPUMSR, I was able to increase the multiplier, but because of
the high core voltage, the CPU would run hot.

The problem with uATX motherboards is reduced features, so the CPU
clock, multiplier and core voltage are not adjustable in the BIOS. I
was able to identify only the Soltek SL-75MIV2 uATX motherboard as
having a BIOS that allows adjusting those values. Unfortunately, the
board seems to be no longer available.

I currently run the mobile XP 2400+ on a full size ATX Aopen AK77-600N
motherboard which allows full BIOS control of the core voltage,
multiplier, and clock frequency.

>Does the BIOS need to be updated specifically for the XP-M?

Yes, except desktop motherboards are not designed to support XP-M,
which means that such a BIOS is not available.

>
>Also, there appear to be different versions of the mobile processors.
>What are the main differences?

Speed and power dissipation.

>
>Thanks,
>Doug

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

In article <gns0911fri2cduqjh9jt19iktqp1gm8sb1@4ax.com>, Andy <1@2.3> wrote:
>On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:
>The BIOS will not recognize the processor, so the voltage and
>multiplier have to be manually set.

Others have said that the processor contains (essentially)
registers which define its correct operating voltage and
multiplier. Thus the BIOS doesn't need to know anything
about the specific processor, as long as it supports the
required voltage and multiplier. Does anyone know for
sure how this is done?

I guess, though, that a given BIOS could just ignore the
voltage/multiplier data from the processor, and simply read
the processor model number and map that into an internal
table to compute the CPU voltage/multiplier, and apparently some
BIOS's do just that. (I wish they'd go back to dip switches ...)

>I used to run a mobile XP 2400+ on
>an MSI KM3M-V uATX motherboard. The default core voltage was set to
>1.7 volts, and there was no way to change it. The default multiplier
>was set to 6, resulting in CPU clock speed of 800MHz. Using the
>program CPUMSR, I was able to increase the multiplier, but because of
>the high core voltage, the CPU would run hot.

What I'm trying to avoid. My system is a low profile and low wattage
system, so I specifically want the XP-m to run at 1.35v.

>The problem with uATX motherboards is reduced features, so the CPU
>clock, multiplier and core voltage are not adjustable in the BIOS. I
>was able to identify only the Soltek SL-75MIV2 uATX motherboard as
>having a BIOS that allows adjusting those values. Unfortunately, the
>board seems to be no longer available.

Yep. I prefer uATX boards because of their smaller size, but there
are too many unnecessary disadvantages - crippled BIOSs, few low
profile cards (actually few low profile cards with low profile brackets).

Thanks again,

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

"dhs" <dhs@cs.utexas.edu> wrote in message
news:d6stgk$39b$1@jib.cs.utexas.edu...
> In article <gns0911fri2cduqjh9jt19iktqp1gm8sb1@4ax.com>, Andy <1@2.3>
wrote:
> >On 20 May 2005 11:15:03 -0500, dhs@cs.utexas.edu (dhs) wrote:
> >The BIOS will not recognize the processor, so the voltage and
> >multiplier have to be manually set.
>
> Others have said that the processor contains (essentially)
> registers which define its correct operating voltage and
> multiplier. Thus the BIOS doesn't need to know anything
> about the specific processor, as long as it supports the
> required voltage and multiplier. Does anyone know for
> sure how this is done?
>
> I guess, though, that a given BIOS could just ignore the
> voltage/multiplier data from the processor, and simply read
> the processor model number and map that into an internal
> table to compute the CPU voltage/multiplier, and apparently some
> BIOS's do just that. (I wish they'd go back to dip switches ...)
>
> >I used to run a mobile XP 2400+ on
> >an MSI KM3M-V uATX motherboard. The default core voltage was set to
> >1.7 volts, and there was no way to change it. The default multiplier
> >was set to 6, resulting in CPU clock speed of 800MHz. Using the
> >program CPUMSR, I was able to increase the multiplier, but because of
> >the high core voltage, the CPU would run hot.
>
> What I'm trying to avoid. My system is a low profile and low wattage
> system, so I specifically want the XP-m to run at 1.35v.
>
> >The problem with uATX motherboards is reduced features, so the CPU
> >clock, multiplier and core voltage are not adjustable in the BIOS. I
> >was able to identify only the Soltek SL-75MIV2 uATX motherboard as
> >having a BIOS that allows adjusting those values. Unfortunately, the
> >board seems to be no longer available.
>
> Yep. I prefer uATX boards because of their smaller size, but there
> are too many unnecessary disadvantages - crippled BIOSs, few low
> profile cards (actually few low profile cards with low profile brackets).
>
> Thanks again,

Have you checked for a newer BIOS version? Sometimes,support for newer cpu's
will be added over time.

 

[Guest]

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

In article <sUMke.247089$cg1.242846@bgtnsc04-news.ops.worldnet.att.net>,
dawg <don't look@worldnet.att.net> wrote:

>"dhs" <dhs@cs.utexas.edu> wrote in message
>news:d6stgk$39b$1@jib.cs.utexas.edu...
>> Yep. I prefer uATX boards because of their smaller size, but there
>> are too many unnecessary disadvantages - crippled BIOSs, few low
>> profile cards (actually few low profile cards with low profile brackets).

>Have you checked for a newer BIOS version? Sometimes,support for newer cpu's
>will be added over time.

Yep, the most recent one was July of last year to support the Sempron.
To their credit, Chaintech has done a good job of supporting this board
(8 BIOS releases over a two year period). I have another uATX board for
which the company (AOpen) stopped releasing BIOS's 3 months after release.
Talk about a short product lifespan.

I may send email to support, but my experience with AOpen for the same
sort of request was pointless. It was like arguing with a rock.
I'd rather AMD devised a scheme for configuring the cpu voltage/multiplier
and bus speeds which doesn't require a new BIOS with every new processor.
Talk about planned obsolescence.