'Performance per watt is the mantra behind Centrino Duo'

-Anand Chandrasekhar, VP & director, Sales and Marketing Group, Intel
Corporation

Intel's recently launched NAPA platform, driven by the
dual core 32-bit Centrino Duo processor, is touted as the big leap in mobile
computing. This processor is expected to outperform the current Sonoma platform
driven by the single core mobile Centrino by 68%, and reduce power consumption
by 28%. Anand Chandrasekhar, vice president & director, Sales and Marketing
Group, Intel Corporation, who was in India for the official launch, spoke
exclusively to Dataquest about the technology and strategic enhancements that
have gone into Centrino Duo.

The
launch of the dual core Centrino Duo processor is touted by Intel as a big leap,
as big as the launch of the Pentium in 1993. How would you explain the
significance of a dual core processor?

Two CPUs can obviously give you maximum efficiency, especially if while
constructing a dual core processor one takes care to look at the thermals and
battery life concerns. A dual core processor is much more efficient in
delivering performance, and this was precisely the reason why Intel decided to
move in the direction of dual core a few years back. The bulk of applications
today, including even common OS like Windows XP that are optimized for Pentium
IV, would run more efficiently on a dual core processor. All of our processors
designed henceforth, for every device including mobile devices, desktops and
servers, will be effectively dual core since it is a much more efficient way of
delivering performance.

Not only would dual
core thus result in significant performance improvement in terms of
applications, it also helps in looking at Moore's Law with a different vision.
According to Moore's law, if processor performance is mapped over a ten-year
period, the last four years would show doubling of performance and next four
years too would show similar doubling. What we are able to do with dual core is
not just double the performance over the next four years by increasing the clock
speed, but achieve so by increasing the number of cores. This is a significant
change since the former, or any similar type of architecture, would have meant
that with doubling of performance thermals too would have gone up. By moving to
a dual core architecture, we have put a lid on this-now with a fixed thermal
envelop we can deliver maximum performance.

AMD has already
launched the 64 bit Turion processors for notebooks. So, what was the rationale
behind launching 32-bit dual core Centrino Duo now, considering that the 64-bit
dual core Merom processors are lined up for 2H 06? How would you react to the
perception that Intel is not vocal on 64-bit processors like AMD?

Intel's product design and development is dictated by what the market
wants. On servers, the market clearly needs 64-bit, and so we went for that.
Ditto with the desktop environment, where too we went for the same architecture
that we are using for servers. In mobile devices like laptops, the market does
not really need 64 bit processors. For 64-bit, one needs to be able to address
4GB of memory. Find me one laptop today that has more than 1GB of memory. The
reason is very simple-it is a question of the battery life. If you add so much
memory into the battery, it will suck because of the additional memory in it. On
a notebook, people care about battery life much more than they do about
computing power.

Intel, therefore,
prioritizes about the battery life and the laptop performance equivalent of
desktop, and, accordingly, we decided that dual core is the right strategy to go
with for now. This does not, however, mean that we are ignoring 64-bit, but that
we will go for it at the right time. And to do it right, we are going to focus
on making sure that battery life does not suffer.

For Centrino Duo to
succeed, it is important to have a proper ecosystem in terms of notebook designs
from various manufacturers as well as applications developed for the technology.
Also, it needs to be fully functional with Vista, the next Windows OS. The
previous version Sonoma did not have adequate support as expected. How are
things different this time?

We would not say that Sonoma lacked in ecosystem support, since when we
launched it a year back there were 150 designs already. However, contrast this
with Centrino Duo and obviously we see tremendous interest in the product, as
there are about 230 designs that we are going to go to market with. To put in
perspective, the first Centrino version was supported by 33 designs. And even if
Centrino Duo is not 64-bit, it will be supporting Vista as and when it is
commercially available. In fact, because of dual core there will be performance
gain in Vista.

One of Intel's key
initiatives in designing mobile processors has been the Common Building Blocks (CBB)
that sought to standardize key components, including displays, drives, batteries
and adapters. How would you explain the rationale behind the CBB program?

The CBB program was started in 2004 when we discovered that there is a cost
bias on a fully manufactured notebook system, in comparison to a desktop. This
cost bias was caused by things which we call “differentiation without a
difference” meaning that people were thinking them as features of the notebook
whereas actually they were not features.  For
example, let us take the standard 14-inch panel in which there is nothing
proprietary about the glass or the connector. Neither one of those added any
features to the notebook that consumers would notice. When we did an analysis,
we found that it was not only with the panel, but also the disk, battery
connectors, optical disk etc, and every manufacturer was involved in it.

It was then that we
started with the CBB program and now you have panels in different notebooks,
which are interchangeable. The differentiation will come when the manufacturer
selects the intensity that they want in the panel since that is a specific user
benefit offered, and the consumer can select that on his own. After the CBB
specs we published, in 2005, many of the components that are CBB compliant
started being manufactured. With Centrino Duo, we are effectively using CBB to
help the notebook market as this program is all about taking off some features
which are not useful. We are Trying to get more ODMs in Taiwan to come within
the CBB ambit.

Dual core thus result in significant performance improvement in terms of applications, it also helps in looking at Moore's Law with a different vision

What has been the
role of Intel's India Development Center in the overall R&D efforts on the
Centrino Duo?

Intel followed a shared design model for almost all the components of the
Centrino Duo platform codenamed NAPA. The processor codename Yonah was given in
Israel, and the development took place in the centers at Yakum in Israel and
Santa Clara, US. The work on the chipset codenamed Calistoga took place in
Bangalore and Folsom, US while that on the wireless module called Golan took
place in Haifa, Israel and San Diego, US.

The assignment given
to the Bangalore team when we started was to become experts in low power to
facilitate battery life from the chipset standpoint. The average power
consumption is about 13 watts but to be able to increase that battery life, we
needed this to be done in 10-11 watts. For that the chipset needed to be much
more intelligent in managing things like display and other components on the
system, as well as power managed.

Prasanto
K Roy & Rajneesh De   

pkr@cybermedia.co.in  

rajneeshd@cybermedia.co.in