Skip to content

NLeSC-COMPAS/kmm

Repository files navigation

KMM


CPU Build Status CUDA Build Status

KMM is a lightweight C++ middleware for accelerated computing, with native support for CUDA.

Highlights of KMM:

  • KMM manages the application memory
    • Allocations of host and device memory is automated
    • Data transfers to and from device are transparently performed when necessary
  • No need to rewrite your kernels
    • KMM can schedule and execute native C++ and CUDA functions

Resources

License

KMM is made available under the terms of the Apache License version 2.0, see the file LICENSE for details.