The difference between Repa and DPH

Question

I recently read a paper about the upcoming generalized stream in the vector and DPH libraries. This seems to be a very interesting development. Now I'm starting to experiment with DPH (starting with GHC 7.6 and plan to upgrade to SIMD version 7.8 when it appears). I also see from the Repa library documentation that it can work with a parallel array. Repa seems to be a mature version compared to DPH , which was not considered ready in terms of GHC 7.4. Now that DPH seems to be ripening, I wonder what the main pros and cons are between the Repa and DPH packages starting with GHC 7.6. I searched StackOverflow and google, but could not find a comparison between Repa and DPH . Hence this question.

Answer 1

From the DPH page on the HaskellWiki:

DPH focuses on irregular parallelism data. For regular parallel data code in Haskell, please consider using the Repa companion library, which is based on the DPH parallel array infrastructure.