It is pointed out that even a single multiplication in the so-called new fast algorithm for circular convolution put forward by Wang in reference[3] is actually not faster than the whole computation for circular convolution in direct multiplication. If th