I want to examine the random numbers generated by a random number generator. Each random number is 4 bytes. I have collected 50,000 numbers, each number is in a separate file (so I have 50,000 files each containing 4 bytes). I'd like diehard to benchmark these random numbers.
One way of feeding my random bytes into dieharder would be to concatenate all 50,000 4 byte values into one file and feed that to dieharder. What I don't understand is: How would dieharder then know that my original numbers were 4 bytes long? For feeding ASCII numbers it supports a header format where one can define "numbit". I didn't find anything like that for binary input. Have I overlooked anything?
Aucun commentaire:
Enregistrer un commentaire