--- Quote Start ---
Hi,
Can someone provide me with the rounding algorithm for 2'scomplement number. rounding opertation (round to nearest) is hard to get for negative numbers in 2's complment. Help me with the solution
sskhati
--- Quote End ---
This example should help (assuming data is 32 bits signed, to be rounded to 16 bits using nearest as defined in Matlab):
lsb <= '0' & data(15);
data_int <= signed(data(32 downto 16)) + signed(lsb);
data_rounded <= data_int(15 downto 0);
you will need to take care of type conversions
alternatively you can add constant '1' to data(32 downto 15) then chop off.
data_int <= signed(data(32 downto 15)) + 1;
data_rounded <= data_int(16 downto 1);
