>> Mr. Needleman: Hi and welcome to
episode number two of the Reporter's Roundtable
CNET's weekly deep dive into one major topic a
week. I'm Rafe Needleman from CNET and today
we're going to be talking about the mobile
computing architecture and how it's evolving
and changing. The big focus on what Intel is
doing. We're heading into the holiday buying
season and most of the laptop vendors have new
lines using mostly Intel CPUs. We'll talk
about these and also the Intel Developer Forum
is coming to San Francisco next week. It's the
biggest Intel chip conference -- one of the
biggest Intel chip conferences of the year and
we'll cover what might be coming up there. We
have two experts this week to guide us through
this. From San Diego Brooke Crothers who is
the author of CNET's Circuit Blog Nanotech and
also a columnist for the New York Times Bits
Section. And from CNET's New York Laptop
Testing Lab senior editor Dan Ackerman. Dan
and Brooke welcome to the show.
>> Mr. Crothers: Thanks.
>> Mr. Needleman: All right. So let's
get started. When I originally started putting
the show together I thought I would focusing on
desktop components because I build a new PC
about every three years and I have to immerse
myself in Intel's new lingo which is always
shifting. But as I talk to you guys I realize
the focus really on innovation in CPUs and the
whole architecture is on the mobile side. So
has innovation really moved from desktop parts
to mobile parts?
>> Mr. Crothers: Well, I would say just
walk into Best Buy and you can get a pretty
good idea immediately of what's going on.
There's like 50 laptops and 10 netbooks and you
look over to the side and there's a few towers.
>> Mr. Needleman: Dusty.
>> Mr. Crothers: Maybe 7 or 8 towers. So
obviously the focus is on mobile and that's
where Intel's focus is on certainly.
>> Mr. Needleman: So looking at the
architecture there a mobile platform is
extremely different from a desktop one. You've
got to focus on many different things. So what
are the things that Intel and other
manufactures are looking at when they're kind
of architecting now for mobile primarily.
>> Mr. Crothers: Well, I would say
>> Mr. Needleman: Uh-huh.
>> Mr. Crothers: And that's really the
theme at IDF. So you make 3 chips into 2 chips
2 chips into 1 chip and you move up the
manufacturing process. Right now we're at 45
nanometer Intel is targeting 32 nanometer. And
all of that gets you a lot of power
efficiently. So for example if you had a
discrete or a separate graphics chip and you
move that on to the CPU and you move that up to
a 32 nanometer process that gets you a lot of
power savings and that's happening in the
mobile space. And that's the focus for Intel's
32 nanometer is in the mobile space. Not so
much -- of course they'll do desktop 32
nanometer but the real focus is on mobile.
>> Mr. Needleman: Okay.
>> Mr. Akerman: And you have to think
about why they want to do that and part of that
is so they can then go into what the end user
end result of that is you can go into the store
and if a machine uses less power they can put a
small battery in it and still say all day
computing. 8 hour computing. Whatever the big
stickers they put on it.
>> Mr. Needleman: Do consumers really
care about batteries? I mean traditionally
I've gone in you know when you go into Best Buy
or Fry's or whatever you see these consumer
laptops and they're gigantic.
>> Mr. Akerman: Well, here's the problem.
All of these laptops are essentially commodity
products and for the most part they have the
same exact components inside. So the only way
you're able the differentiate from the
competition is through it's kind of design I
guess also price and then you know any kind of
special features whether they're real or
imagined that you can put on a big sticker on
the box or put in you know the Best Buy
>> Mr. Needleman: So -- sorry, Brooke.
>> Mr. Crothers: No go ahead.
>> Mr. Needleman: So how are Intel and
some of the other players that put the parts in
these or make the parts for these computers how
are they kind of pushing things along? I mean
we've talked about the die size or process
sizes are getting smaller. They're going to 32
nanometer. What else are they pushing along
into these products to make them you know more
mobile as opposed to just you know portable
desktops but actual mobile devices? What are
they pushing on the manufactures and then on to
>> Mr. Crothers: I would say look at
netbooks and look at this knew category
ultra-thins. So netbooks is all about power
savings, right? And it's a very small item
it's a very you know -- to put it in simplistic
terms a small chip and it uses very little
power. So my take is that battery life is
important now. It wasn't that important before
but now it's really important. So you look at
all the now ultra-thins that came out and I
think Dan can probably validate this it's all
about you know they're saying 8 hour 10 hour
battery life and that's really different then
before where it was just like you know giving
us maybe 3 or 4 maybe 5 hours max. So you know
all the technologies we just addressed this a
minute ago really that enable that you know
longer battery life. Because think about it.
When you take a netbook or a laptop with you
now you know it's going to be really crucial
that you get -- if you go to an all-day
conference I mean it's going to be really
important that you get all-day battery life.
And I don't think people expected that before
but I think that expectation is there now.
>> Mr. Akerman: Just very quickly on
battery life. What they used to do is say, Oh,
yeah, 8 hour battery all-day battery with a big
asterisk next to it. And then if you read the
fine print it said you know if you turn the
monitor down to like 20 percent brightness and
you turn off the wifi and you don't so this and
you as long as you don't open any programs. Now at least
they can make the same number claims but
they're actually inching back towards reality.
There are a lot of very fine for instance
fairly inexpensive netbooks that have 6-cell
batteries that don't stick out too badly from
the back and they actually will run for 5 1/2,
6 hours on some strenuous stuff like video
>> Mr. Needleman: So let's talk about I
want to learn about Atom because of course that
is -- traditionally a netbook is a small
computer powered by an Atom chip. What is the
Atom chip? How is a different from your
standard Core Duo or Core 2Duo or whatever?
What makes it special?
>> Mr. Crothers: Well, it's not very
sophisticated. Let's put it that way. It's a
pretty simple chip and if you look at the Core
2Duo and the upcoming Nehalem it's a very super scaler
and all that stuff. Atom is not -- though I
guess some chip people may take issue with
this -- but it's not really a super scaler
chip. And super scaler has been the design MO
for the last ten years. So it's a really simple
chip and thereby gets you the power savings.
And that's the simplest way to put it.
>> Mr. Akerman: So it's kind of a dumbed
down chip on a modern process?
>> Mr. Crothers: Yes. It's a very modern
process. It's very cheap to make. And so in
other words you can get a lot of Atoms on a wafer.
So what Intel's thinking is we want -- here's
this new business model where we have to make
you know we have to make chips for cheap
computers so we want to have a profit margin on
this chip. So put a lot of chips on the wafer
and you know therefore you get better yields.
Theoretically a decent profit margin. However,
I would take issue with that. And Intel is
sort of -- I wrote a blog about this and
actually written about this a number of times
that it created a Frankenstein monster of sorts
in that it's becoming too successful because Atom
was originally designed for the mid-market for
the mobile internet device. Think of it as a
high-end smart phone and all of a sudden you
know [in audible] and Acers chimed in and
everyone started making netbooks all of a
sudden. And Intel did not expect this. And
they will tell you that they did but they
>> Mr. Akerman: They became a victim of
their own success in a lot of ways.
>> Mr. Crothers: Yeah. Yeah. Exactly.
>> Mr. Needleman: So now people in the
chat room are asking did I just say that all
netbooks are Atoms. Let me be clear about
that. My understand was that initially people
were trying to say that a netbook is an Atom
powered machine although I think that has
changed now and netbooks are just whatever
people want to call them. And we've got
netbooks and we've got netops which are
basically netbooks in desktop form factors and
I mean is this Atom chip line going to continue
as the standard you know mobile chip, the
traditional mobile chip becomes more efficient
are we going to see the death of the Atom?
>> Mr. Akerman: No. I don't know we
will. But remember the very first netbooks
actually had Celeron chips in them the Atoms didn't
come until some time later. And that's still I
would 90 plus percent of all netbooks. A
couple of people have tried to put other chips
in netbook Avia had their nano-processor I
think I saw like one Samsung system ever and
AMD had their Athlon Neo which they kind of
pitched as a step up from the Atom. Something
that was a little bit more powerful they could
charge a little more for which is kind of the
holy grail of everyone who is kind of stuck in
this netbook moras where they can't sell a
machine for more than 299 or 399 now. Which is
what I meant when I said they were victims of
their own success. But the problem with that
Athlon Neo is and the reason why I've only ever
seen it in one system is it just really wasn't
that much better even though they tried to
charge another 300 bucks for the final product.
>> Mr. Needleman: There is getting a
little far afield here but I'm worried
about the netbook or the laptop market in general
because of netbooks being so low cost. At 300
bucks or 250 or even if you go done even
>> Mr. Akerman: And they will.
>> Mr. Needleman: I don't understand how
anybody is making any margin.
>> Mr. Akerman: They're not.
>> Mr. Needleman: The Windows license fee
is 50 bucks itself.
>> Mr. Akerman: And they might have to
knock that back too which is something I think
has been hinted at recently. I'll just say
very briefly on that subject. A point I've tried
to make repeatedly to people in the industry
who may not want to hear this is that with the
advent of netbooks and Atoms the consumer has
finally woke up and realized that for years
they've been buying too much computer. And now
they understand that. And now they have a lot
more power in the marketplace than when they
just went into a store didn't know what they
wanted and somebody said here you need to buy
this $1000 system.
>> Mr. Needleman: What other important
things is Intel or the other manufactures doing
to kind of balance the need for you know
processing power and energy consumption there?
>> Mr. Akerman: I think they're pushing
these new consumer ULV chips which are a much
different animal than the old ULV chips you
used to find in those little like 12 inch we
used to call them ultra-portable systems that
were like you know $1,500, $2000 that now they
call this knew middle ground CULV. The first
examples of it however -- and I'm not sure how
Brooke feels about this -- have not been super
impressive to me.
>> Mr. Needleman: Are these CPUs or
support chips you're talking about?
>> Mr. Akerman: There's are CPUs we're
>> Mr. Needleman: Okay.
>> Mr. Crothers: Right. So what Intel is
trying to do is -- so Sean Maloney who is sort of
considered to be the next -- may be the next
CEO of Intel. And I talked to him about this a
few months ago and he was saying you know
netbooks are great and we're glad the netbook
category is successful but we have this knew
chip called they call it the CULV now it's
called the ultra-thin category. And that's
really their strategy to get back the profit
margins. So I agree. It's a real challenge
for Intel. The netbook has been too successful
and I really don't think they can stop it. And
bear in mind that at IDF Intel will be
announcing a new Atom platform called Pinetrail
which is the biggest probably leap for Atom in a long
time because Atom you know -- did you notice
this last time they upgraded Atom they went
from what 1.6 gigahertz to 1.66 gigahertz. It
wasn't a huge leap. So obviously they're
trying to control that category so it doesn't
get too high performance. But Pinetrail is
going to actually put the graphics right on the
CPU die so that will be interesting to see if
that gets you more power efficiency and better
>> Mr. Akerman: But real goal there is to
lock out the nVidia ion which is nVidia's
graphics chip for low powered system and
they're really not interested in helping you
out by putting the graphics on the chip they
just want to lock out the competition and make
you buy their full package rather than just the
CPU and then use your own you know chip set
>> Mr. Needleman: Let's talk about
graphics for a second and extend on that. One
of the things we're seeing now on mid and
high-end notebooks is dual graphic systems on
the same product. Lenovo and Apple both have
machines and I think Sony does as well and
probably some others with switchable graphic
subsystems. You can go integrated which is the
one that's built into the mother board or you
can go the discrete graphic which is I don't
know how they solder them in or socket them in
or whatever. Is that the wave of the future or
are consumers going to need less and less of
the discrete graphics?
>> Mr. Akerman: They definitely already
need less of the discrete graphics. I've talked to a
bunch of manufactures about this very topic
this week because everyone is come around
showing their holiday lines. And I think a big
problem is most people what who have a machine
with switchable graphics don't even realize
they have it. It's almost like this instant on
prelaunch operating systems. They don't even
know they're there. If you look at the Apple
Macbook pro there's a setting somewhere in the
menu that says more battery or more
performance. It doesn't actually tell you what
>> Mr. Crothers: Well, yes. That's a
good question. I think and you know nVidia is
not going to go away let's put it that way.
nVidia is a big company and they deliver really
good graphics and they have a pretty good
marketing machine. It's not up to Intel's
level but you know I think they make a point
that for example as Dan was just talking about
in the netbook space their point is if you want
you know let's say laptop level performance in
a netbook you should go with our ion platform.
And that is of course it's on a discrete -- it's
not on a dicrete chip set, excuse me,
a discrete graphics chip it's
actually integrated into the nVidia chip set.
Just for argument sake let's say it is a
discrete chip because it is a fairly
high-preference chip so their argument is you
know you do need a discreet chip to get real
mainstream PC level performance in a netbook.
So you can pay you know let's say $400 or $350 for
a netbook and get mainstream PC level
performance. So the point being that yeah it
is necessary. I'm you do need discrete
graphics level performance because Intel hasn't
really delivered that yet. I mean Intel
integrated graphics is let's face it. It's
okay but it's not great.
>> Mr. Akerman: You know it's okay if you
need the computer to turn on and that's about
>> Mr. Needleman: Oh I don't know. I use
integrated graphics. I don't do -- aside from
Google Earth I don't do anything that's
>> Mr. Akerman: But here's the test of
the netbook and the reason why people are
finally realizing they need something other
than just the default Atom with the integrated
graphics is try to go to HULU and try to watch
what they call HD which is their 480p stream
and 9 times out of 10 your netbook is not going
to -- it's just going to stutter.
>> Mr. Needleman: Okay. Point taken.
And it surprises me actually that Intel is in
that position. I mean AMD bought ATI shouldn't
they be ahead of the game on this? You know
why are they not ahead of the game here?
>> Mr. Crothers: I can address that.
They do have better graphics than Intel. But
they really have problems with their CPU.
Their mobile CPU. I'm not talking about their
desktop I'm not talking about their servers.
They really have a problem with it's called
Turion and they have a thing called Neo. Neo
is okay. It's being marketed more or less as a
netbook platform but it's still pretty much
Atom level preference. And Turion has never
really caught fire. And Intel has just
completely dominated the mobile space. And I
bring this up with AMD often. Typically when
you get PR from AMD it's all about desktops,
yawn, and servers which are of course very
important that's a very high profit margin
segment for them. But you hear very little
about the mobile space. These days you're
hearing a little bit about Neo but you know for
AMD it's a real challenge. Here's this huge
huge market that's just growing leaps and bounds
everyone's buying laptops and they're not a big
player and they do have ATI so if you go to
Best Buy you can get an AMD laptop with an ATI
and it's better graphics than Intel graphics
but you know the average consumer they go to
the Intel solution. So that's a problem.
>> Mr. Needleman: That's brand success.
>> Mr. Akerman: I mean we went to some
retail stores for a big back to school round up
for laptops. They get all the models off the
retail store shelfs and test them. And you're
entirely right. You can get the AMD version a
little bit cheaper than the Intel version in
very similar systems but in our actual
benchmark testing -- and we're talking the 499,
550, 599 systems -- there's really a big
preference difference and that's enough for you
to say I'm going to spend 50 bucks more and get
the Intel version instead. And the problem
with that Neo that Brooke mentioned and again
the reason why it's really only ever been in
one laptop I think a second one from HP is
finally going to come out but I think it going
to be a dual core Neo is instead of AMD saying
hey the Atom is successful Let's make our own
clone of that and try to capitalize on that
they said, Let's make our own more expense
clone of that and try to up-sell it when it
really wasn't that much better or enough to
make you pay more for it.
>> Mr. Needleman: Let's move on to the
last tech -- whoa I need a new chip -- the last
tech topic before we get on to the IDF here and
talk about mobile -- and talk about
communications rather. Is this going to be the
season of WiMAX? And if not when are we going
to see that?
>> Mr. Crothers: Well, that's a good
question. I think WiMAX is a tough call
because it's not as unsuccessful as people
think but it's certainly not that successful
yet. I think it takes time. I was reading an
article on a Japanese web site I think it was
yesterday. And it was about a Thinkpad that
had a WiMAX you know one of the few laptops
that has WiMAX built in and the writer was
complaining and again this is in Japan which
supposedly is you know fairly far you know
relatively advanced in WiMAX infrastructure.
And the whole article was about him complaining
he couldn't get a good connection in downtown
Tokyo. So that's a problem for WiMAX. But you know it
takes time for these technologies. You have to
build up the infrastructure. And I'm not going
to say that you know WiMAX is toast yet.
>> Mr. Needleman: Too early to bury it,
huh? But we're not going to see it this year?
>> Mr. Akerman: Oh, no. Definitely not.
I mean what's the one test market? Is it
Denver? Is that what it was?
>> Mr. Needleman: I think so.
>> Mr. Crothers: [inaudible].
>> Mr. Needleman: Not here of course.
>> Mr. Akerman: You have the
infrastructure problem. You got LTE at the
same time and you have again who is the market
for this? People are finally everyone finally
has wifi on their laptops now. And people
finally have routers in their houses and they
finally have wifi in all the coffee shops and
municipal wifi in the parks. Are we just going
to tear that down and start again? Who's going
to pay for all of this?
>> Mr. Needleman: You are. That's the
answer. All of us aren't we?
>> Mr. Akerman: And we're also try to
up-sell people to getting their own personal 3D
subscription data connections at the same time.
We spent years building that market up and
that's finally becoming almost an acceptable
thing to get when you buy a laptop and want the antenna built in.
>> Mr. Needleman: We're final seeing
these subsidized netbooks. You know, so --
>> Mr. Akerman: Yes. We're starting to.
A handful of cases although the subsidies
aren't nearly enough. You know they will sell
it to you for 199 instead of 299 with a two
year contract. It really should be 99 bucks,
49 bucks, free even. You get a 299 Aces 2d50
that's you know or a Dell mini 10v those are
299 netbooks. If you're going to get 60 bucks
a month from me for two years just give it to
me for free.
>> Mr. Needleman: Yeah. No kidding.
Hey, let's close with looking at the Intel
Developer Forum that's coming up in San
Francisco next week. I'll be down at the demo
start-up conference. Brooke's coming up here
to go to IDF. What do we expect to see here?
How important a show is this?
>> Mr. Crothers: Very important. And
I'll throw the chip heads out there I'll give you some code
names and you know code names are confusing but
they are part and parcel of IDF. It's a big
part of IDF. So what we have is we have first
we have this thing called Clarksfield which is
a core which Intel Nehalem type architecture.
They're calling it commercially the core I3,
I4, I5, I7. And so the chips that I'm going to
list are mostly core series chips. The
Clarksfield is a 45 nanometer quad core Nehalem
mobile processor. It's the first Nehalem
mobile processer and that will be a big thing at
IDF. Then there's this thing Clarkdale. It
gets confusing really fast here. Clarksfield.
Clarkdale is a 32 nanometer Nehalem desktop chip
integrating graphics on the CPU for the first
time. And there's this thing call Arrandale
which is a next generation sort of after
Clarksfield which is a 32 nanometer Nahalem
mobile chip also integrating graphics on the
CPU. So as you can see integrating graphics on
the CPU is a big thing for Intel. There's this
thing called moristown [assumed spelling] which
everyone I think it's a pretty well-known code name. It's
a 32 nanometer system on a chip. And so a
system on a chip is obviously just what it
means everything is on one chip basically. And
that's going to be Intel's -- that's 32
nanometers as I said. And that's going to
Intel's vehicle to get into the smart phone
states. And that's where the arm versus Intel
arm being the Qualcomms, the TIs, the free
scales of the world. That's Intel and I think
this is going to be one of the big topics at
IDF. What Intel is trying to do is they're
trying to downsize into the smart phone. And
that's not easy to get x86 which you know is in
servers and high-end gating machines. And it's
a power you know inherently a power hungry
architecture. You get that in a smart phone
that is no mean task. And then you have arm
the arm camp trying to up-size into things
called smart books. So that's an interesting
thing to watch. And just a couple more things
the Pinetrail thing which is the Atom chip that
I mentioned. Then Laraby [assumed spelling] is
Intel's first discrete graphics chip in about
ten years. And that's going to be very
interesting because that's going to compete
with nVidia and ATI.
>> Mr. Akerman: Ten years?
>> Mr. Crothers: Yeah. They had a
discrete graphics chip way back when that no
one has ever heard of. If you know Intel
you've heard of it and if you've followed the chip
industry you've heard of it. But that in fat didn't
pan out. So this is their first discrete being
operative discrete graphics chip in about ten
years I think. May be a little more than ten
>> Mr. Needleman: Wow. All right.
Brooke, thanks for dialing in. Dan, thanks as
well. I tell you something this show went on a
little longer than I planned but I learned a
lot and I'm going to keep doing this round
table just because it's incredibly educational
for me and I hope everybody else is enjoying it
as well. Thanks guys again for dialing in
Skyping in whatever it was. Everybody thanks
for listening to Reporter's Roundtable. We are
life each Friday at 1:00 Pacific time at
live.cnet.com. Next time we're going to have
our first outside the CNET infrastructure
guest, Aaron Patzer. He's the CEO of Mint or
now some VP at Intuit because the company Mint
is a personal finance web site was acquired by
Intuit for $170 million. He'll be here in the
studio to talk about what is going to happen
with personal finance and Intuit and Mint. And
Eleanor Mills our security reporter will also
be here to help me grill Aaron. Digital City,
Dan Akerman's podcast is moving from the spot
on Friday to the primo spot of Mondays 3:00
Eastern time Noon Pacific. This is about
technology for the urban environment. You can
find Brooke Crothers on CNET on the Nanotech
blog. And if you have question about this or
other podcasts you can reach me at
firstname.lastname@example.org. I would love to hear your
feedback on the show and the show format. And
that is it for Reporter's Round Table Number 2.
Thanks everyone for listening.