Using function cvecvmltf I found incorrect results. Please find attached the example source files. When the program is run it first prints the results from cvecvmltf and then those from the definition of a complex multiplication. The second results are correct.
There are two data files: arrayA.txt and arrayB.txt each with 128*2 elements. In the example I have used N=16, but every N<=128 will do.
The results from cvecvmltf are incorrect both on the simulator and on the 21469 EZ-Kit Lite board.
A very remarkable event occurs with the statement int maxval = 0; maxval is never used and can be removed. When I do so all results are correct!!
I hope this phenomenon also appears with your tests.