https://stackoverflow.com/questions/68857974/how-to-perform-sorting-in-numba-cuda-kernels-without-using-local-cuda-array (Edit: now deleted and refactored into other stack exchange questions, see below)
Also posting it here. Hopefully gets some more Google visibility in the future for others to benefit from by being on stackoverflow.com. It links to a lot of other resources related to cuda.local.array.
1 Like
Reworked into the following two questions: