Using Quasirandom algorithm as a randomness test?

This pages extend more the UDLR rotor router extension (case3 of the Main page) that presents no moiré.

The rotor router with CED extension as a clean border (like already known for the original rotor router algorithm)

Here we use 200K iterations in the algorithm.

If at each iteration we initialise the agent at center position using the pseudorandom generator embedded in C code in

the UDLR rotor router with the extension to set an empty cell to the current value of the agent at center position, we observe some unclean borders in the circle.

Does it means that the pseudorandom generator of the C code is not perfect?

On the left: agent initialized with pseudorandom "rand" of C code. agentCurrentValue=rand()%4;

On the right: agent initilized with an UDLR loop. (U=1,D=2,L=3,R=4)

details of the initialisation loop: (please refer to the main page for full C code)

// init starting value for agent

if(agentCurrentValue < 4)

{

agentCurrentValue++;

}

else

{

agentCurrentValue=1;

}