The NIDRU performs many calculations during runtime. The precision of these calculations is controlled by the MASK_CG attribute, which permits the ability to trade off between precision and resource usage. For example:
•MASK_CG = 1111111111111111 sets calculations to internal 16-bit precision (highest precision, highest resource usage)
•MASK_CG = 1111111111111110 sets calculations to internal 15-bit precision (lower precision, lower resource usage)
TIP: Lowering the internal precision to 10-bit has a negligible impact on the performance.
Reducing the internal calculation precision negatively impacts the jitter tolerance. The amount has to be evaluated either in hardware or in simulation.
S_MAX is the expected maximum number of extracted samples per clock cycle. NMAX is the maximum number of bits processed per clock cycle.
IMPORTANT: It is mandatory to set S_MAX > N_MAX. N_MAX is calculated using the following equation:
N_MAX = truncate (fDIN/fREFCLK+1)
S_MAX_EYE is an internal parameter and must be set according to the following equation:
Equation 6