SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Texas Instruments - Good buy now or should we wait?
TXN 159.33-1.8%Nov 14 9:30 AM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: DOUG H who wrote (4535)7/6/1999 9:54:00 AM
From: slacker711  Read Replies (1) of 6180
 
A comparision between DSP's and ASIC's for 3G basestations....note that the authors work for TI.

csdmag.com

CSD 2019

3G Basestations and DSPs

By Jean-Pierre Demange and David Smalley

One of the foremost goals of IMT-2000 (the ITU's road map for the so-called third-generation of wireless technology) is to
implement new capabilities ranging from high-speed Internet access to video, with wireline quality. The challenge to
basestation manufacturers is to find ways to implement these features as well as to support multiple air interface standards.
And, they'll have to ensure all of these capabilities for less money, at lower operating costs, in less space, and in less time.

One of the primary considerations is how to rapidly implement new signal processing functions without resorting to expensive,
time-consuming hardware redesigns. There are currently two schools of thought about this. In the first, digital signal processing
is employed through the development stage, with the design migrating to an ASIC in the end product. In the second, some of
the functions of the end product are handled by ASICs and others by digital signal processors (DSPs). The appeal of the
second approach lies in the ability of DSP-based designs to be reprogrammed in the field, allowing new standards and
features to be implemented through software without the hardware redesign and deployment required by ASIC solutions.

The inherent flexibility of digital signal processing, and its ability to rapidly implement modifications through software, will make
DSPs more attractive, even to designers who have traditionally relied heavily on ASICs. This is not to say that ASICs have no
place in tomorrow's basestations. Their place is assured for the foreseeable future, but their role will surely change. DSPs and
ASICs will complement each other, performing tasks to which each is best suited.

DSPs will implement more functions than before in basestation transceivers for several reasons. First, their clock rates will
exceed 500 MHz, providing more MIPS per dollar and MIPS per square mm of board space. Second, most DSPs have
open architectures that allow easy third-party development of end-user applications, as well as smart antennas and other
infrastructure-specific enhancements. DSPs will also evolve to incorporate more wireless-specific instructions and adjunct
coprocessor building blocks to optimize performance.

The overwhelming need for rapid time-to-market will preclude service providers from employing expensive, time-consuming
hardware modifications to accommodate new standards or features. Manufacturers will get better results in a fraction of the
time by adding features to a DSP-based solution through software, rather than taking an ASIC for another run through the
foundry.

Service providers are unlikely to tolerate the design philosophy of any basestation manufacturer that requires more time or
more money than is deemed necessary to incorporate features such as Internet access and video. This isn't surprising given
that competitors are trying to pare away their subscriber base with the promise of some new feature or perceived advantage.

The future synergy of basestations and DSPs seems assured — assuming the DSP community rises to the challenge with a
continuing stream of advances. While past performance is hardly a guarantee of future performance, DSP performance has
already advanced by a factor of forty-two in less than a decade. Even the most sober industry pundits believe a
factor-of-twenty performance increase is likely within the next five years.

When the market calls for Internet access, full-motion video, and other features, DSPs will likely be the preferred way of
implementing them. That's good news for service providers that need to offer these capabilities quickly, for the manufacturers
responsible for building the basestations, and for consumers waiting to use them.

Jean-Pierre Demange is a wireless infrastructure business development manager in Texas Instruments' France location.
He has an MS in electronics from Montpellier, France University. He can be reached at jp-demange@ti.com.

David Smalley is a senior member of the technical staff for wireless DSP applications in Texas Instruments' Norcross,
Georgia office. He has a BEE and MSEE from the Georgia Institute of Technology. He can be reached at
d_smalley@ti.com.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext