Intel talks about new Napa platform

  • Thread starter Thread starter Yousuf Khan
  • Start date Start date
Australian IT - Intel touts new laptop chip (Scott Hillis in San
Francisco, DECEMBER 14, 2005)
http://australianit.news.com.au/articles/0,7204,17565420^15321^^nbv^,00.html

Doesn't work for me but no matter: according to the quote you have below,
which also appears here
http://yahoo.reuters.com/financeQuo...tfh43917_2005-12-14_00-39-30_n13362603_newsml
it's just a Reuters article.
Albeit, the article is not perfect:

Yeah, I guess what an anal...yst doesn't know just doesn't exist. Who are
those people and how do they hold down a job?
 
Yeah, I guess what an anal...yst doesn't know just doesn't exist. Who are
those people and how do they hold down a job?

In order to climb the corporate ladder, it is paramount to have
excellent brownnosing skills. The rest is optional.

NNN
 
Australian IT - Intel touts new laptop chip (Scott Hillis in San
Francisco, DECEMBER 14, 2005)
http://australianit.news.com.au/articles/0,7204,17565420^15321^^nbv^,00.html

Albeit, the article is not perfect:

Hmm, I dunno who else complained but it's been corrected at Yahoo.Reuters:
http://yahoo.reuters.com/financeQuo...tfh65192_2005-12-14_18-44-53_n13362603_newsml

"...... and has one-upped Intel by being the first to launch a 64-bit
mobile processor."

I guess "trumped" would have been too strong for Intel to stomach as an
established fact.:-)
 
George said:
Hmm, I dunno who else complained but it's been corrected at Yahoo.Reuters:
http://yahoo.reuters.com/financeQuo...tfh65192_2005-12-14_18-44-53_n13362603_newsml

It might have been me. I'm just too impatient. :-)
"...... and has one-upped Intel by being the first to launch a 64-bit
mobile processor."

I guess "trumped" would have been too strong for Intel to stomach as an
established fact.:-)

But there were other hard-to-stomach quotes in there about Intel:
"Napa will make Intel a more capable competitor against AMD's products. AMD is having a modicum of success in mobile," said Nathan Brookwood, head of semiconductor research firm Insight 64.

Brookwood saying "will make Intel a more capable competitor against AMD"
is just a kick in the balls isn't it?

Yousuf Khan
 
It might have been me. I'm just too impatient. :-)


But there were other hard-to-stomach quotes in there about Intel:


Brookwood saying "will make Intel a more capable competitor against AMD"
is just a kick in the balls isn't it?

Err, just a modicum of one.:-)
 
It might have been me. I'm just too impatient. :-)


But there were other hard-to-stomach quotes in there about Intel:


Brookwood saying "will make Intel a more capable competitor against AMD"
is just a kick in the balls isn't it?

Yousuf Khan

Or a reference to
http://www.channelregister.co.uk/2005/12/15/amd_yamato_centrino/
AMD unwraps notebook reference platform

On the surface of it, Napa/Yonah doesn't look anymore such a clear-cut
winner. If Pentium M scales up with the number of cores the way P!!!
did (I see no better reference, since PM is derived from good old P6
core - PPro->P!!! - and has very little in common with netbust Xeons
of today), and Turion64 scales like A64X2, Napa would have quite some
catching up to do in performance department. And if the AMD target of
5 hours of battery life is reached, that will make them competitive in
thin-n-light ultra-mobility dept., now occupied exclusively by
Centrino.

NNN
 
Or a reference to
http://www.channelregister.co.uk/2005/12/15/amd_yamato_centrino/
AMD unwraps notebook reference platform

On the surface of it, Napa/Yonah doesn't look anymore such a clear-cut
winner. If Pentium M scales up with the number of cores the way P!!!
did (I see no better reference, since PM is derived from good old P6
core - PPro->P!!! - and has very little in common with netbust Xeons
of today),

One of the key differences here is that the Pentium-M DOES share the
same bus as the Xeon, which offers significantly more bandwidth than
the old PIII bus. As such, it should scale to dual-core much better
than an old PIII might.

Of course, the Athlon64 should scale even better to dual-cores due to
it's integrated memory controller and hypertransport design.
and Turion64 scales like A64X2, Napa would have quite some
catching up to do in performance department. And if the AMD target of
5 hours of battery life is reached, that will make them competitive in
thin-n-light ultra-mobility dept., now occupied exclusively by
Centrino.

I really don't like those "battery hour" figures when used in
reference to CPUs. Usually the highest power consumer of a modern
laptop is the display, not the CPU. Hard drives, memory and video
chips all figure in to the equation for power consumption. On the
other side of things is the battery size, which varies widely from one
notebook to another. I'd much rather the processor makes talk about
actual watts of their chips.

Still, I don't think that the dual-core Turion's will be entirely out
of firing range of the dual-core "Yonah" Pentium-M chips. They should
be at least able to get down into the "ML" range of 35W (vs. ~25-30W
for the Pentium-M). Of course, there is also the chipsets to
consider, and I understand that mobile chipsets for AMD processors
(from nVidia and ATI) currently consume more power than those from
Intel (if anyone has any numbers here, please feel free to share
them!).
 
Tony said:
One of the key differences here is that the Pentium-M DOES share the
same bus as the Xeon, which offers significantly more bandwidth than
the old PIII bus. As such, it should scale to dual-core much better
than an old PIII might.

Plus, unlike the kludged-up P4 DC's, I think the Yonahs will be using an
internal communications channel, instead of going out over the FSB.
I really don't like those "battery hour" figures when used in
reference to CPUs. Usually the highest power consumer of a modern
laptop is the display, not the CPU. Hard drives, memory and video
chips all figure in to the equation for power consumption. On the
other side of things is the battery size, which varies widely from one
notebook to another. I'd much rather the processor makes talk about
actual watts of their chips.

An interesting observation was made about the relative battery lives of
Turion vs. Pentium-M: Turions conserve more energy under average
workloads, while P-M conserves more energy under heavy load. The only
reason I can think of this difference is that both chips will use no
more than its rated maximum energy consumption under heavy loads, but
usually P-M has a lower consumption rating than typical Turion (typical
Turions are usually an ML @ 35W, rather than the more trifty MT @ 25W).

But under average loads, their automatic power management is keeping
them well under maximum consumption ratings. So it's possible that the
AMD power management is doing a better job of keeping Turion under its
maximum power consumption than the P-M's power manager is.
Still, I don't think that the dual-core Turion's will be entirely out
of firing range of the dual-core "Yonah" Pentium-M chips. They should
be at least able to get down into the "ML" range of 35W (vs. ~25-30W
for the Pentium-M). Of course, there is also the chipsets to
consider, and I understand that mobile chipsets for AMD processors
(from nVidia and ATI) currently consume more power than those from
Intel (if anyone has any numbers here, please feel free to share
them!).

I have no figures.

Yousuf Khan
 
An interesting observation was made about the relative battery lives of
Turion vs. Pentium-M: Turions conserve more energy under average
workloads, while P-M conserves more energy under heavy load. The only
reason I can think of this difference is that both chips will use no
more than its rated maximum energy consumption under heavy loads, but
usually P-M has a lower consumption rating than typical Turion (typical
Turions are usually an ML @ 35W, rather than the more trifty MT @ 25W).

But under average loads, their automatic power management is keeping
them well under maximum consumption ratings. So it's possible that the
AMD power management is doing a better job of keeping Turion under its
maximum power consumption than the P-M's power manager is.

What is the absolute performance of the said chip at said full load?
Does the Turion/P-M has a higher peak performance or similar?

Similarly, at average load, which is faster?
 
The said:
What is the absolute performance of the said chip at said full load?
Does the Turion/P-M has a higher peak performance or similar?

Similarly, at average load, which is faster?

I think at full load, all they're looking for is whether the CPU usage
shows 100%. In average load, they're probably looking for anything that
keeps cpu utilization below 100% (preferably below 50% most likely).
They're not really measuring performance of the apps at that level, just
what the battery life is.

Yousuf Khan
 
I think at full load, all they're looking for is whether the CPU usage
shows 100%. In average load, they're probably looking for anything that
keeps cpu utilization below 100% (preferably below 50% most likely).
They're not really measuring performance of the apps at that level, just
what the battery life is.

In that case, it isn't a very useful indicator is it? After all, if I
can do what I need in say 80% of the time, even tho the battery life
at that kind of load is only 90% the competition, I'm still getting
around 10% more battery life no?
 
In that case, it isn't a very useful indicator is it? After all, if I
can do what I need in say 80% of the time, even tho the battery life
at that kind of load is only 90% the competition, I'm still getting
around 10% more battery life no?

True enough, though I think for the vast majority of laptop users it
isn't going to be much of an issue. From my own personal experience,
and that of most people I see, the only time they REALLY stress their
system enough to make a smallish (~20%) difference in performance
noticeable is when they have it plugged in to the wall. Most of us
don't carry a laptop around to do a large rendering project or a
complex Place 'n Route run, we carry the laptop around to show people
the results of said projects.
 
The said:
In that case, it isn't a very useful indicator is it? After all, if I
can do what I need in say 80% of the time, even tho the battery life
at that kind of load is only 90% the competition, I'm still getting
around 10% more battery life no?

Uh, I'm not quite sure what you just said. So I'll just take a best
guess. :-)

There was test of the Dothan vs. Turion recently. It looks like how they
do the less than heavy load tests is to do a regular applications test,
and then to put a specific amount of idle time between the tests to
simulate a laptop environment. This is the test where Turion does better
than Pentium M.
Our original hypothesis about Turion was that it would be more power efficient under idle usage circumstances, like the Life test in Winstone's BatteryMark test. Since the Life test consists of running common productivity applications with timed periods of idle-ness to simulate true user interaction, Turion's architecture lends itself to being more efficient.

http://www.laptoplogic.com/resources/detail.php?id=17&page=16

Yousuf Khan
 
Back
Top