migration of cudpp 1.1 radix sort

Accelerated Computing CUDA CUDA Programming and Performance
1

hi all…

Does anyone know if there is a timeline for migrating the CUDPP 1.1 radix sort (which supports key/value) to OpenCL?

thx

2

I would like to know this too. Been using the bitonic sort that shipped with the SDK so far, but the limitations on list sizes are troubling.

3

It’s fairly easy to modify the NVidia opencl radix sort example to support values. Look at the values operations in CUDPP and migrate to oclRadixSort:

CUDPP 1.1 oclRadixSort

radixSortBlock() radixSortBlockKeysOnly()
reorderData() reorderDataKeysOnly()

In both cases the values are shuffled in the same manner as the keys.