Northwood vs Prescott

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

I've heard that newer Prescotts run cooler, and would be comparable to
Northwood processors. Is this correct? What advantages does the
Prescott offer over Northwood Hyperthreading processors?

 

[Paul]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <odj531hlvktu9lcpg810ol8d5iv1kc13c5@4ax.com>, _R
<_R@nomail.org> wrote:

> I've heard that newer Prescotts run cooler, and would be comparable to
> Northwood processors. Is this correct? What advantages does the
> Prescott offer over Northwood Hyperthreading processors?

Well, the Prescott can be overclocked to 4+GHZ. That has to
be worth something.

As for power savings, it depends. The C1E feature, slows the
processor down to the same speed as a 2.8Ghz processor. If the
processor is, in fact, a 2.8Ghz, there is zero savings. C1E savings
are more appreciable for faster processors.

I expect you've read one too many glowing Intel reports on Tomwhardware.
I wonder how much he gets paid for writing those...

If you want real numbers, the datasheets can be downloaded from Intel,
and you can compare numbers for the various running states.

As for benchmarks, they are +/- . What I mean by that, is some
benches are faster and some are slower. The Prescott has a longer
pipeline, and if code has a lot of branches (especially incorrectly
predicted branches), a lot of results are thrown away. The
Northbridge pipeline is a bit shorter, so is not affected as much.

On the cache side of things, the Prescott was given a larger cache.
If a particular code is cache sensitive, the Prescott will win over
the Northwood.

If you do decide to get a Prescott, I would recommend a good third
party heatsink/fan and some thermal paste. An XP-120 or a Zalman
7000 or 7700, will give you a possibly quieter cooling solution.
They might also put less mechanical stress on the motherboard. It is
important to keep the Prescott below 70C while it is running 100%, at
whatever overclock you are using, because it will throttle back the
processing rate if it gets too hot. And your third party heatsink is
only as good as the case ventilation around it. You have to move
a good deal of air past the processor area, to get rid of the
"warm cloud".

Paul

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Prescotts run significantly hotter and need good cooling. I would suggest
that you make sure you have good airflow in your case. My Northwood 2.6
overclocked to 3.0 would idle at 32C. At work, I have a Prescott 3.0, it
idles at 40C. Both have a stock Intel cooler installed, both cases are in
open enough areas that there is good aiflow around them, both cases have the
side air duct recommmended by Intel and both cases have a 120mm low rpm fan
at the rear so the comparison is pretty accurate in my mind.

"_R" <_R@nomail.org> wrote in message
news😱dj531hlvktu9lcpg810ol8d5iv1kc13c5@4ax.com...
> I've heard that newer Prescotts run cooler, and would be comparable to
> Northwood processors. Is this correct? What advantages does the
> Prescott offer over Northwood Hyperthreading processors?
>

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Sat, 12 Mar 2005 07:16:20 -0500, nospam@needed.com (Paul) wrote:
>
>I expect you've read one too many glowing Intel reports on Tomwhardware.
>I wonder how much he gets paid for writing those...

?? Actually I haven't. Does Tom's Hardware favor the Northwood, or
are you advocating AMD? I do some sensitive software development
(often in ASM) and need to make sure that the software is Intel-
compatible. For home systems I might try AMD. Or if they had a CPU
that was a *lot* faster than Intel's.

[good info snipped...thanks, Paul]

>If you do decide to get a Prescott, I would recommend a good third
>party heatsink/fan and some thermal paste.
> Paul

That's what I was concerned about. I don't often overclock anything.
The requirement for 3rd party heatsink would make sense for someone
who has to wring the last Hz out of their CPU, but I usually don't
worry about it.

 

[Paul]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <dv47315dvajp0b2dfpf1h1odcaou4f6cm2@4ax.com>, _R
<_R@nomail.org> wrote:

> On Sat, 12 Mar 2005 07:16:20 -0500, nospam@needed.com (Paul) wrote:
> >
> >I expect you've read one too many glowing Intel reports on Tomwhardware.
> >I wonder how much he gets paid for writing those...
>
> ?? Actually I haven't. Does Tom's Hardware favor the Northwood, or
> are you advocating AMD? I do some sensitive software development
> (often in ASM) and need to make sure that the software is Intel-
> compatible. For home systems I might try AMD. Or if they had a CPU
> that was a *lot* faster than Intel's.
>
> [good info snipped...thanks, Paul]
>
> >If you do decide to get a Prescott, I would recommend a good third
> >party heatsink/fan and some thermal paste.
> > Paul
>
> That's what I was concerned about. I don't often overclock anything.
> The requirement for 3rd party heatsink would make sense for someone
> who has to wring the last Hz out of their CPU, but I usually don't
> worry about it.

My personal opinion, is Tomshardware should wait until Intel does
something worthwhile, and then congratulate them. The only unbiased
comparison that counts, is the power values listed in the datasheet,
because they aren't subject to editorializing.

Intel should be congratulated on the PentiumM. Truly noteworthy TDP.
Intel is getting better on halt-type power ratings. At least
the 3.8Ghz is in the same ballpark as the Northwoods (really
depends on what the VID value is, which is not stated). Full
power is still on the high side.

PentiumM Model 770 2.13GHz FSB533 2MB cache TDP=27 watts

P4EE LGA775 3.73GHz FSB1066 PRB=1 2MB cache 90nm TDP=115 watts
Stop grant 62 amps at 1.2-1.4V, 74 watts or more
The 3.6GHz 660 model is Enhanced auto halt 40 amps
P4EE LGA775 3.46Ghz FSB1066 2MB cache 0.13u TDP=111 watts
Stop grant 40 amps at 1.525-1.6V, 61Watts or more
P4 LGA775 3.8GHz FSB800 1MB cache 90nm TDP=115 watts
Enhanced auto halt 37 amps at 1.2-1.425V, 44Watts or more
P4 LGA775 3.8F FSB800 1MB cache 90nm TDP=115 watts
Stop grant 56 amps at 1.2-1.425V, 67Watts or more

Northwoods S478
P4 3.4Ghz/FSB800/512KB TDP=89 Watts
Stop grant 27 amps at roughly 1.5V, 40.5Watts
P4 3.4Ghz/FSB800/2MB TDP=103 Watts
Stop grant 32 amps at roughly 1.5V, 48Watts

AMD 4000+ 2.4Ghz 1MB cache TDP=89W
Min P-state 22Watts StopGrant 10.6Watts
AMD FX55 2.6GHz 1MB cache TDP=104W
Min P-state 25Watts StopGrant not specified

My comment about the 3rd party heatsink, has more to do with noise
than with cooling. If the intended use for the computer is
in an un-airconditioned room, the Intel fan will wail in the
summertime. It likely won't be nearly as bad in an airconditioned
office setting. Third party solutions give you the opportunity
to drop the noise level.

As for the AMD versus Intel thing, I have all Intel processors,
except for an AthlonXP. I'm quite impressed by its thermal
characteristics. I rate systems by how uncomfortably hot they
make my computing room in summer.

Paul

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Sat, 12 Mar 2005 11:56:58 -0500, "zzipper" <zzipper@sympatico.ca>
wrote:

>Prescotts run significantly hotter and need good cooling. I would suggest
>that you make sure you have good airflow in your case. My Northwood 2.6
>overclocked to 3.0 would idle at 32C. At work, I have a Prescott 3.0, it
>idles at 40C. Both have a stock Intel cooler installed, both cases are in
>open enough areas that there is good aiflow around them, both cases have the
>side air duct recommmended by Intel and both cases have a 120mm low rpm fan
>at the rear so the comparison is pretty accurate in my mind.
>

I wonder if you or Paul would be kind enought to answer a question for
me.

I have a P4C800E-Deluxe running the 2.8C (Northwood)
(no over-clocking) in an Antec P-160 case with an Enermax
EG475P-VE-SFMA-470W-SATA power supply. No additional cooling has been
added other than the Intel stock cooling which came with the CPU.

MotherBoard Monitor (MBM) indicates the CPU temp hovers around 35C at
idle. Since I do a lot of DVD processing, MBM shows the CPU temps
reach 52C for as long as an hour when the CPU is at full throttle as
MPEG files are being encoded and processed. It's never gone over the
52C mark.

From what I've read on this forum, I'm given to understand 52C is
nothing to worry about but I would appreciate your thoughts if you can
take a moment.

The temp sensor on the Antec P-160 does indicate a somewhat higher
idle temp (43C) but never seems to go beyond 52C at full CPU use.

Can you or Paul share any insight?

Thanks

f

 

[Paul]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <r1s831d5lutea18068nkeqh31dl96tcjmg@4ax.com>,
nospam4me@notaol.com.invalid wrote:

> On Sat, 12 Mar 2005 11:56:58 -0500, "zzipper" <zzipper@sympatico.ca>
> wrote:
>
> >Prescotts run significantly hotter and need good cooling. I would suggest
> >that you make sure you have good airflow in your case. My Northwood 2.6
> >overclocked to 3.0 would idle at 32C. At work, I have a Prescott 3.0, it
> >idles at 40C. Both have a stock Intel cooler installed, both cases are in
> >open enough areas that there is good aiflow around them, both cases have the
> >side air duct recommmended by Intel and both cases have a 120mm low rpm fan
> >at the rear so the comparison is pretty accurate in my mind.
> >
>
> I wonder if you or Paul would be kind enought to answer a question for
> me.
>
> I have a P4C800E-Deluxe running the 2.8C (Northwood)
> (no over-clocking) in an Antec P-160 case with an Enermax
> EG475P-VE-SFMA-470W-SATA power supply. No additional cooling has been
> added other than the Intel stock cooling which came with the CPU.
>
> MotherBoard Monitor (MBM) indicates the CPU temp hovers around 35C at
> idle. Since I do a lot of DVD processing, MBM shows the CPU temps
> reach 52C for as long as an hour when the CPU is at full throttle as
> MPEG files are being encoded and processed. It's never gone over the
> 52C mark.
>
> From what I've read on this forum, I'm given to understand 52C is
> nothing to worry about but I would appreciate your thoughts if you can
> take a moment.
>
> The temp sensor on the Antec P-160 does indicate a somewhat higher
> idle temp (43C) but never seems to go beyond 52C at full CPU use.
>
> Can you or Paul share any insight?
>
> Thanks
>
> f

There are three temperatures to consider - the room temperature,
the case ambient temperature, and the CPU temperature. Each one
influences the other, so cooling performance is measured by
taking the difference between these three temps. (I cannot find
a link now, but if memory serves, I think AMD may have recommended
no more than a 7C delta between case ambient and room air temp.)

I have a similar configuration - P4 2.8C, P4C800-E, Antec Sonata.
The Sonata has the same internals as the P160, but a different
external design.

What i noticed with my case, is the 120mm fan is starved for air.
When I open the door on the side of the case, the fan speed drops
due to the air being able to move more freely. There are not enough
inlet vents on the front of the case, for the capacity of the fan.
On the Sonata, this is easy to fix. There are plastic pieces on the
front of the computer, that can be removed. I get to keep the upper
door, while exposing more vent area in the lower half of the case.
By doing that, more room air blows immediately over the drives in
the lower front drive frame.

I think I got about 8C reduction in case ambient, which allows the
CPU to cool down by an equal amount.

52C is not too much for the processor, so there is no issue there.
I would be more concerned about the case ambient and the effect
this could have on disk drive life. Hard drives are sensitive
to the combination of high humidity and high temperature. The
IBM web site used to have manuals for their disk drives, and there
are curves in a graph that show acceptable levels of humidity
and temp. For example, at 60% relative humidity, the drive
case temp is supposed to be maintained below 37C. 60% R.H. is
the point at which the carpets in your room will start to
mildew.

In an air-conditioned room, the R.H. will be around 40%, so
both temp and humidity are improved at the same time. That is
why someone working in an AC environment has little to worry
about.

I had an air conditioner failure, and worked for about a month
at temps and humidities around the level listed above. I lost
a Maxtor drive (in one of my other computers) during that time.
Statistically speaking, that doesn't mean anything, but I have
to wonder if it wasn't more than a coincidence.

If you make any improvements to your computer case, remember to
pamper the disk drives, because they are more sensitive to
temp than your processor. Adding a duct and vent hole to the
case side panel, helps the processor, and will allow the case
air temp to drop, but a better solution, is to increase the vent
area in front of the disk drives, so the fresh air moves past
them first.

HTH,
Paul

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Sun, 13 Mar 2005 23:22:37 -0500, nospam@needed.com (Paul) wrote:

>
>There are three temperatures to consider - the room temperature,
>the case ambient temperature, and the CPU temperature. Each one
>influences the other, so cooling performance is measured by
>taking the difference between these three temps. (I cannot find
>a link now, but if memory serves, I think AMD may have recommended
>no more than a 7C delta between case ambient and room air temp.)
>
>I have a similar configuration - P4 2.8C, P4C800-E, Antec Sonata.
>The Sonata has the same internals as the P160, but a different
>external design.
>
>What i noticed with my case, is the 120mm fan is starved for air.
>When I open the door on the side of the case, the fan speed drops
>due to the air being able to move more freely. There are not enough
>inlet vents on the front of the case, for the capacity of the fan.
>On the Sonata, this is easy to fix. There are plastic pieces on the
>front of the computer, that can be removed. I get to keep the upper
>door, while exposing more vent area in the lower half of the case.
>By doing that, more room air blows immediately over the drives in
>the lower front drive frame.
>
>I think I got about 8C reduction in case ambient, which allows the
>CPU to cool down by an equal amount.
>
>52C is not too much for the processor, so there is no issue there.
>I would be more concerned about the case ambient and the effect
>this could have on disk drive life. Hard drives are sensitive
>to the combination of high humidity and high temperature. The
>IBM web site used to have manuals for their disk drives, and there
>are curves in a graph that show acceptable levels of humidity
>and temp. For example, at 60% relative humidity, the drive
>case temp is supposed to be maintained below 37C. 60% R.H. is
>the point at which the carpets in your room will start to
>mildew.
>
>In an air-conditioned room, the R.H. will be around 40%, so
>both temp and humidity are improved at the same time. That is
>why someone working in an AC environment has little to worry
>about.
>
>I had an air conditioner failure, and worked for about a month
>at temps and humidities around the level listed above. I lost
>a Maxtor drive (in one of my other computers) during that time.
>Statistically speaking, that doesn't mean anything, but I have
>to wonder if it wasn't more than a coincidence.
>
>If you make any improvements to your computer case, remember to
>pamper the disk drives, because they are more sensitive to
>temp than your processor. Adding a duct and vent hole to the
>case side panel, helps the processor, and will allow the case
>air temp to drop, but a better solution, is to increase the vent
>area in front of the disk drives, so the fresh air moves past
>them first.
>
>HTH,
> Paul

Thank you for taking the time to respond, and for providing such
detailed (and very useful) information.

I have always been aware that ambient temperature affects the temps
inside the case and even saw it yesterday as I was composing my
earlier post. At 80 degrees, yesterday was a much warmer day than it
has been thus far this year. That 20 degree rise in temperature was
certainly reflected in the temps recorded by MBM and the case
monitoring.

I've also had a concern for the 3 hard drives I have installed. The
Antec P-160 has three large front-side slots which allows external air
to pass directly over the hard drives on its way to the processor
before exhausting out the back. But bear in mind that I process a lot
of video. An MPEG video file the type of which I process is generally
in the 6 to 8 gigabyte range. I typically move video data between two
(non-system) drives as a means of speeding these processes. (i.e. a
movie being encoded on my E drive will save that encoded file to the F
drive.) The encoding and authoring software may use less than 30 % of
the processor, but the disc activity of files moving from one drive to
the other means the drives are running for as much as two hours while
these large files are being processed and saved. Because there's so
much strain on these drives, your points are very well taken and I
thank you for them.

f