compare performance a new Rand object vs a lock & using the same one

[mayabar]

50K iterations of both options
Step 1 - one thread
Step 2 - multiple threads

[mayabar]

Tested for different number of parallel threads (5 and 10), different delays between calling the random number generation function (0, 50, 100 millis), running 1000 iterations in each test.
Conclusion: creating random source once + using a mutex is more efficient than creating random source of each random value generation.