Hey there.
Has anyone succeeded in saving the Mersenne Twister, provided in the SDK, current state so that the next call to the kernel produces random data without repeating the values from the random sequence since its beginning? The kernel provided initializes the RNG on each thread but it’s not prepared to return to the CPU and then be called again to continue generating data. It will simply reset and start all over.
Muchas gracias
Hey there.
Has anyone succeeded in saving the Mersenne Twister, provided in the SDK, current state so that the next call to the kernel produces random data without repeating the values from the random sequence since its beginning? The kernel provided initializes the RNG on each thread but it’s not prepared to return to the CPU and then be called again to continue generating data. It will simply reset and start all over.
Muchas gracias