Autocorrelation Test

When I started programming microcomputers I found a repeated need for random numbers. Dave Dodson suggested implementations and tests of their quality, including autocorrelation, which I implemented by brute force on the CDC 6500.

I'd been introduced to the notion of testing random sequences in an undergraduate programming class. I had implemented Conditional Distribution then.

I'd also had an elementary introduction to systems at various points in my EE education including computing Impulse Response via Convolution in an EE lab.

I'd been programming in CDC assembly language but not yet done much floating point work outside of Fortran. I chose to code the inner loop in assembler which I called from Fortran.

I coded up a Tausworthe Generator and fed progressively longer sequences into my autocorrelation. The results looked good.

The computation ran fast enough that I decided to try the full sequence. This was terrible, of course, since the generator repeats and is thus perfectly correlated.