EETimes India - December 16, 2008 - (Page 6) In Focus | DSPs Design, debug with DSPs By Aseem Vasudev DSP Applications Manager Analog Devices India Pvt Ltd Digital signal processors are the central processing units of today’s sophisticated electronics systems. Especially with some of the DSP architectures being designed and optimised for signal processing and microcontroller applications, DSPs are now widely used in almost all kinds of electronics systems. Various advanced and intensive designs such as automotive systems, medical systems, surveillance, military, multimedia products, and handheld digital devices like PDAs and smart phones use DSPs. Processor selection DSPs are basically classified as fixed point DSPs and floating point DSPs. This is mainly based on the data type of the operand as being fixed (16/32 bit integer or 1.15/1.31 format fractional) or IEEE-32-bit floating point. Selection depends on the algorithm that operates on the data. Some algorithms require bit accuracy like certain lossless decoders and encoders and these systems therefore use DSPs that support fixed-point (16/32-bit) arithmetic. Some others require higher dynamic range in which case the obvious choice is a DSP that supports floating-point arithmetic natively. Once this basic requirement is determined, several other factors need to be considered when selecting a DSP. The most vital of these are MIPS (million instructions per second), MFLOPS (mega floating-point operations per second), the availability of onchip controllers and peripherals, the amount of on-chip memory, and power consumption. MIPS and MFLOPS imply the speed of operation that helps determine whether the algorithm could be executed in real-time or not. Onchip controllers and peripherals allow seamless, point-to-point interfaces without the need to use external glue-logic to interface Figure 1: Multiprocessing using multiple DSPs. Figure 2: A single core with two processing elements. with external devices such as SDRAM, DDR-I/II, data converters, and industry standard protocols such as USB, UART, CAN, and LAN. Fundamentally, having such onchip controllers and peripherals reduces the total system cost. Similarly, power consumption has become a major consideration especially for battery-operated devices that need less powerhungry DSPs. Dynamic power management has thus become an important and must-have feature for DSPs used in such products. Multiprocessing approaches Certain applications either require faster processing speeds or more processing elements. For example, there are situations where a single application may require more MIPS than what a processor can support, or multiple instances of the same or different applications to be run on multiple DSPs. This is called as multi-processing. Multiprocessing is commonly seen in applications such as medical imaging systems, military applications like RADAR, and some multi-channel audio applications like the high definition audio in AVRs and professional audio sound cards. Based on the application(s) requirement, it can be achieved Discuss FPGAs outperform DSPs? For highly parallel algorithms, FPGAs and other custom hardware offer better performance and greater cost reduction than DSPs. Do you agree? with one of the following approaches: • Using multiple DSPs—This approach uses multiple processors to perform sequential multi-processing or parallel multi-processing (figure 1). In sequential multi-processcontinued on page 6 EE Times-India | December 16-31, 2008 | www.eetindia.com http://www.embeddeddesignindia.co.in/SEARCH/SUMMARY/technical-articles/signal processing.HTM?ClickFromNewsletter_081216 http://www.eetindia.co.in/SEARCH/SUMMARY/technical-articles/%22smart phone%22.HTM?ClickFromNewsletter_081216 http://www.eetindia.co.in/SEARCH/SUMMARY/technical-articles/%22smart phone%22.HTM?ClickFromNewsletter_081216 http://www.embeddeddesignindia.co.in/SEARCH/SUMMARY/technical-articles/algorithm.HTM?ClickFromNewsletter_081216 http://www.embeddeddesignindia.co.in/SEARCH/SUMMARY/technical-articles/algorithm.HTM?ClickFromNewsletter_081216 http://forum.embeddeddesignindia.co.in/FORUM_POST_1000039251_1200096594_0.HTM?ClickFromNewsletter_081216 http://www.embeddeddesignindia.co.in/SEARCH/SUMMARY/technical-articles/MIPS.HTM?ClickFromNewsletter_081216 http://forum.embeddeddesignindia.co.in/FORUM_POST_1000039251_1200096594_0.HTM?ClickFromNewsletter_081216 http://forum.embeddeddesignindia.co.in/FORUM_POST_1000039251_1200096594_0.HTM?ClickFromNewsletter_081216 http://www.eetindia.com/STATIC/REDIRECT/Newsletter_081216_EETI02.htm?ClickFromNewsletter_081216
Table of Contents Feed for the Digital Edition of EETimes India - December 16, 2008 EETimes India - December 16, 2008 Contents National Semiconductor Data Plane Processing Challenges DSP Design, Debug with DSPs TES 2008, NCMIPMV '08, ICACT '08, VLSI Conferene 2009, ICETiC 2009 EETimes India - December 16, 2008 EETimes India - December 16, 2008 - Contents (Page 1) EETimes India - December 16, 2008 - National Semiconductor (Page 2) EETimes India - December 16, 2008 - Data Plane Processing Challenges DSP (Page 3) EETimes India - December 16, 2008 - Data Plane Processing Challenges DSP (Page 4) EETimes India - December 16, 2008 - Data Plane Processing Challenges DSP (Page 5) EETimes India - December 16, 2008 - Design, Debug with DSPs (Page 6) EETimes India - December 16, 2008 - Design, Debug with DSPs (Page 7) EETimes India - December 16, 2008 - Design, Debug with DSPs (Page 8) EETimes India - December 16, 2008 - TES 2008, NCMIPMV '08, ICACT '08, VLSI Conferene 2009, ICETiC 2009 (Page 9) EETimes India - December 16, 2008 - TES 2008, NCMIPMV '08, ICACT '08, VLSI Conferene 2009, ICETiC 2009 (Page 10)
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.