Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
It dependes on your application. usually FPGAs are used in higher performance applications(say faster speed applications). But they are not flexibla as DSPs.
If you need answer of given question, especially of the question as global as you have asked it is alyaws better to geve as much information as posible.
So if you need more adequate information just describe what are you going to do and you will get the answer.
which one is better??
a BULDOZER or a TOYOTA pick up !
It dependends for what ..
Both come in different FLAVORS ..
so many things MATTER .. to your QUESTION
0) HOW MUCH PROCESSING POWER DO YOU NEED
Do you need a BULDOZER or a SHOVEL ?
1) Is the application SIGNAL PROCESSING ?
(then you can try to compare )
2) Do you need FLOATING POINT CALCULATIONS
Current DSPs have one two MAC (multiply accumulator) units. These units are used sequentially. If one needs more than two MACs (for example, over 100 tap FIR filter with sample rate of 200MHz) then parallel MACs with single cycle computation is possible to realize only using FPGAs with current trends.
It depend s on the no of MAC operations - for simple algorithms DSPs are ok .IN adaptive array signal processing whrer the MAC operations are many , the FPGA performs 10 ,to 100 times faster .
The usual technique is that the Data Flow Graph is directly
mapped into FPGA. And the resulting structure implements
the algorithm fully in parallel.
Dotzens of hardware multipliers in modern FPGAs support it.
None DSP microprocessor could do it.
Therefore the rule of thumb is:
FPGA is useful when the real time sampling frequence is higher than 1-10 MHz;
DSP microprocessor is useful when the sampling frequence is less than 1-10 MHz.
I recommended FPGA as better solution.
Really, hundreads of DSP microprocessors could support the
GHz data flow, sharing the common input data buffer.
But in this situation the FPGA based solution will really
occupy less hardware and consume less energy.
DSP microprocessor is out of questions for applications
with very heavy algorithms like CELP voice coding.
Because there is not any useful technology to map such
an algorithm into FPGA.
May be Celoxica and Handel C?
But it seems to be rather stiff because
the programmer has to parallel the algorithm by hand.
But it is the question of the behavioral synthesis.
FPGA is not hard to use. YOU'd better to use it to design interface circuits and control circuits.
DSP is good at digital signal processing,i.e FFT,IFFT...
FPGA's have more parallel units and do reegular tasks faster such as
image processing but
DSP's are better for tasks that change a lot like signal processing
I guess the answer depends on the application and the product. If you are simulation and developing systolic based filtering using integers, an FPGA is probably a better choice for cascaded design. But if you are doing floating point number crunching algos and would like the ease of development thru bundled software libraries that came with your device software toolset, a DSP such as TMC320 series would be great! Keep in mind none is a complete replacement to other for the particular need and that's why both devices exist and thrive in the industry.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
