This is a cross post of @drifter1/programming-cuda-atomic-functions-and-synchronization by 
Another proof that parellelization doesn't always improve the results and should be always used with caution! :)
Read next
From @drifter1
This is a cross post of @drifter1/programming-cuda-atomic-functions-and-synchronization by
Another proof that parellelization doesn't always improve the results and should be always used with caution! :)