AFTBX BENCHMARKS

AFTBX vs LOOP: HOW FAST ?

This plot shows the average time different computers take to run 1 iteration and 200 realizations of an adaptive filter of length M = 100. AFTBX is, on average, 8.6 times faster than loop implementations!

AFTBX benchmark

ARE THE AFTBX RESULTS ACCURATE ?

The following figure compares the output of the AFTBX with that of classical loop implementations. The differences are normalized by the largest intermediate result and compared with MATLAB's precision: eps(). AFTBX yields the exact same result than loop implementations!

AFTBX benchmark
GO BACK AND DOWNLOAD THE AFTBX