Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Branch: master
Commits on Apr 23, 2012
  1. Add a README

    authored
  2. @mchakravarty
  3. @mchakravarty
  4. @mchakravarty
  5. @mchakravarty
  6. @mchakravarty

    Submodule accelerate-cuda

    mchakravarty authored
  7. @mchakravarty
  8. @mchakravarty

    Clean up after merging

    mchakravarty authored
  9. @mchakravarty
  10. @mchakravarty
  11. @mchakravarty
  12. @mchakravarty
  13. @mchakravarty
  14. @mchakravarty

    Update accelerate-io

    mchakravarty authored
  15. @mchakravarty

    README updates

    mchakravarty authored
  16. @mchakravarty
Commits on Apr 22, 2012
  1. @mchakravarty
  2. @mchakravarty

    Submodule accelerate-io

    mchakravarty authored
  3. @mchakravarty
  4. @mchakravarty

    Fix merge-related wibbles

    mchakravarty authored
  5. @tmcdonell @mchakravarty

    fix dimensionality reification for foldSeg

    tmcdonell authored mchakravarty committed
  6. @tmcdonell @mchakravarty

    export Sugar.showShape

    tmcdonell authored mchakravarty committed
  7. @tmcdonell @mchakravarty

    improve 'fromList' implementation

    tmcdonell authored mchakravarty committed
    traverse the list once writing data instead of using (!!) repeatedly. also report an actual error if the input list is too short, rather than "(!!) out of bounds".
  8. @tmcdonell @mchakravarty

    fix fold implementation for zero-sized arrays

    tmcdonell authored mchakravarty committed
  9. @tmcdonell @mchakravarty
  10. @tmcdonell @mchakravarty

    parameterise Segments to be of integral type

    tmcdonell authored mchakravarty committed
    this is important for hardware implementations, as native Haskell Int support could entail 64-bit integers which might be expensive or impossible
    
    Conflicts:
    
    	Data/Array/Accelerate/Array/Sugar.hs
    	Data/Array/Accelerate/Prelude.hs
  11. @tmcdonell @mchakravarty
  12. @tmcdonell @mchakravarty

    wibbles

    tmcdonell authored mchakravarty committed
  13. @tmcdonell @mchakravarty

    migrate quickcheck related to accelerate-examples

    tmcdonell authored mchakravarty committed
  14. @tmcdonell @mchakravarty

    use boolToInt instead of explicit branch

    tmcdonell authored mchakravarty committed
  15. @tmcdonell @mchakravarty

    fix finaliser implementation (honestly, this time)

    tmcdonell authored mchakravarty committed
    attach the finaliser to the ArrayData instead of the underlying UArray, which doesn't work in ghci
  16. @tmcdonell @mchakravarty

    fix kernel invocation containing tuple types

    tmcdonell authored mchakravarty committed
    CUDA 4.0 changed the way kernel parameters are passed, attempting to match the given parameters with the real kernel arguments. This doesn't work in our setup, since we use struct types to cheat in code generation so skeletons are parametric over dimensionality and array type. I new version of the bindings package provides the necessary behaviour again.
  17. @tmcdonell @mchakravarty

    improve debug message handling/init

    tmcdonell authored mchakravarty committed
Something went wrong with that request. Please try again.