Aspects of present disclosure relate to random number generator, a method and a computer program product of improving entropy quality of the random number generator. The method may include: receiving, at an input/output interface module of the random number generator, a request to generate a random number having a predetermined number of random bits, and starting a random bit generating loop to generate each of the random bits of the random number to be generated. In certain embodiments, random bit generating loop may include: incorporating a CPU Time as a randomness factor in generating random number to improve entropy quality, including non-deterministic memory-subsystem latencies in entropy extraction, such as those introduced by unpredictable cache movements, generating a Candidate Bit by using a Clock Time, and generating a random bit for random number by using a von Neumann unbiasing analysis module, until every random bits of the random number is generated.