I have an array of float values.
I want to sort it so that I get a new array with indices. For example:
source array => array indexes
[0.99, 0.72, 0.56, 0.88, 0.11, 0.50] => [0,3,1,2,5,4]
Perhaps there is a ready-made implementation of cuda kernel, but I have not been able to find it yet. The thrust is not suitable for me due to the fact that the data is in a batch and I cannot mix the sort results, and serial execution is a bottleneck.