Coding timinig relationship

[vb_thecapt]

Hi everybody! I have a question about coding&timing/delay relationship.
Considering a code like this

for index 0 to MAX_N loop
varMatrix:= CONV_INTEGER(matrix(index));
diff := a-varMatrix;
if(diff < VAL) then
c:= coeff(diff);
else
c := 1;
end if;
prod := ker(index)*c;
...
end loop;

where i create a variable for every accessed array element , and considering this code

for index 0 to MAX_N loop
if( abs(a-CONV_INTEGER(matrix(index))) < VAL) then
prod := ker(index)*coeff( a-CONV_INTEGER(matrix(index)) );
else
prod := ker(index);
end if;
...
end loop;

where i use every time that i need the "whole" indexed-array element.
If these two codes have different hardware mapping, which of these two codes are faster?
Code 1 implies a register (thus a delay) for every created variable?
Code 2 implies only that a "wire" is taken from a source?
Sorry if it sounds a bit messy

 

[joris]

Intermediate variables don't imply registers, so both codes should be exactly equivalent.

Variables only imply registers when their values are being read while they haven't previously been set in the same clock tick.