Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Fetching contributors…

Cannot retrieve contributors at this time

84 lines (64 sloc) 2.411 kb
Name: meta-par-accelerate
Version: 0.3.5
Synopsis: Support for integrated Accelerate computations within Meta-par.
-- Version history:
--
-- 0.3: Initial release.
-- 0.3.1: tweak dependencies
-- 0.3.2: tweak resulting from abstract-par-accelerate-0.3.2
-- 0.3.3: Remove erroneous ConstraintKinds usage.
-- 0.3.4: Egad, ditto.
-- 0.3.5: Remove dep on accelerate-io. Set default to newAccelerate. Reactivate real CUDA.
Description:
This package provides a 'Control.Monad.Par.Meta.Resource' for building meta-par
(<hackage.haskell.org/package/meta-par>) schedulers with GPU support.
This package also provides a complete scheduler for CPU plus GPU
execution. It supports the 'Control.Monad.Par.Par' monad
programming model with additional support for GPUs.
Homepage: https://github.com/simonmar/monad-par
License: BSD3
License-file: LICENSE
Author: Ryan Newton, Adam Foltzer 2011-2012
Maintainer: Ryan Newton <rrnewton@gmail.com>
Copyright: (c) Adam Foltzer 2011-2012
Stability: Experimental
Category: Parallelism
Build-type: Simple
Cabal-version: >=1.8
Library
Exposed-modules:
-- This is a complete scheduler for shared-memory
-- (parallel) and GPU computation:
Control.Monad.Par.Meta.AccSMP,
-- This is a resource used to build schedulers:
Control.Monad.Par.Meta.Resources.Accelerate
Build-depends:
base >= 4 && < 5
, abstract-par
, meta-par
, abstract-deque
, abstract-par-accelerate
-- , abstract-par-offchip
, QuickCheck
, transformers
, vector
, array
ghc-options: -Wall
if flag(cuda)
cpp-options: -DACCELERATE_CUDA_BACKEND
if flag(newAccelerate) && flag(cuda)
Build-depends: accelerate-cuda
if flag(debug)
cpp-options: -DDEBUG
if flag(newAccelerate)
Build-depends: accelerate >= 0.12
-- , accelerate-io >= 0.12
else
Build-depends: accelerate >= 0.10 && < 0.12
flag newAccelerate
default: True
-- Sets the default backend to CUDA:
flag cuda
default: False
flag debug
default: False
Jump to Line
Something went wrong with that request. Please try again.