Hi,
I am currently working on a circuit which has to perform Hamming distance computation between large bit vectors (>500 bits).
I was surprised not to find much information on how to implement this type of operation *efficiently* on FPGA technology.
So far I have been investigating two approaches (combining tables and counter for the bitcouting part). I observed that the choice of table size (3 or 4 address bits) had a significant impact (20%) on the area cost of the operator.
I feel that there are many subtle trade-offs in such implementations, and I was wondering if anybody had been looking at this problem (most of the articles I stumbled accross dealt with the correcting code issue, rather than focusing on the Hamming distance realization in itself).
Thanks in advances for the input.
Regards,
Steven