fastest parallel sorting algorithm

Hey,

can any one tell me which is the best available parallel sorting algorithm in terms of speed?

Regards,
sagar

This paper was recently posted by someone else in the forum:

http://www.cs.chalmers.se/~dcs/TechReports/gpuqsort.pdf

It describes a CUDA-specific version of quicksort, which they show performs very favorably compared to other parallel sorting algorithms.