[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ecc] Decoding different (N,K) combinations




How about just going through the whole thing. Are you sure that the chien
algorithm is the bottle neck of your system, maybe you can just waste a few
clocks and nothing will happen...

Lior



                                                                                                                   
                    Nimrod Mesika                                                                                  
                    <nimrodm@beze        To:     [email protected]                                                 
                    qint.net>            cc:                                                                       
                    Sent by:             Subject:     [ecc] Decoding different (N,K) combinations                  
                    owner-ecc@ope                                                                                  
                    ncores.org                                                                                     
                                                                                                                   
                                                                                                                   
                    06/16/2001                                                                                     
                    11:14 AM                                                                                       
                    Please                                                                                         
                    respond to                                                                                     
                    ecc                                                                                            
                                                                                                                   
                                                                                                                   




I'm having trouble implementing the Chien search algorithm for
shortened codewords (i.e., N<255 for GF(256)).

It seems like I can do one of the following:

o Start the search from the last byte of the codeword (the lowest
  power of alpha, alpha^(-1)) and proceed up to alpha^(-N). Thus,
  the output will be in reverse order with respect to the input of
  the decoder.

o Start the search from alpha^(-N) and proceed to alpha^(-1). This
  is simple if N is constant. However, if N may be different for
  different codewords, another set of (non-constant) multipliers are
  necessary so that coefficients of lambda(x) and omega(x) will be
  multiplied by alpha^(-N) before the search starts.

  Of course, this method also requires a ROM table of alpha^(-N)
  for every possible N.


I was wondering if there is a simpler method? Anyone?

I'll be happy to provide my C code and (almost finished) verilog
code.

--
Nimrod.

--
To unsubscribe from ecc mailing list please visit
http://www.opencores.org/mailinglists.shtml




--
To unsubscribe from ecc mailing list please visit http://www.opencores.org/mailinglists.shtml