Memory space and execution model
TNL has separate concepts for the memory space and execution model, which are represented by different template parameters. See the Core concepts page for details.
- Most other libraries have separate types for CPU and GPU data structures (e.g.
Vector and cuVector):
- These libraries have the concept of a "memory space" which is configurable as a template parameter:
- CUV
- CUSP - but CUSP uses Thrust, so
device_memory might be the same as host_memory if OpenMP is used as the device
- Kokkos - they have a concept of a "memory space" and "execution space", but there is also some default choice of the spaces, possibly even through command-line arguments (in which case the array type would be polymorphic, because something has to store the current memory/execution space)
- These libraries have transparent access to the data from GPU and CPU:
- the CUDA toolkit itself, via
cudaMallocManaged
- cudarrays - they have custom virtual memory system using
cudaMalloc and the standard host allocator
- These libraries select the (default) device based on some macro (this approach is way too simple, because multiple different devices cannot be combined):
- These libraries do not abstract memory space, only execution model:
- RAJA
- Nebo (also with a macro-based selection)
Multidimensional arrays
TODO: compare the implementation of multidimensional arrays (features described in the merge request: https://gitlab.com/tnl-project/tnl/merge_requests/18 )
- http://cpptruths.blogspot.cz/2011/10/multi-dimensional-arrays-in-c11.html
- http://www.nongnu.org/tensors/ (last commit in 2012)
- https://bitbucket.org/wlandry/ftensor/src
- Eigen tensors - Many operations, expression templates, either pure-static or pure-dynamic sizes, only column-major format (row-major support is incomplete), little GPU support.
- cudarrays - Only up to 3D arrays, both static and dynamic, compile-time permutations using
std::tuple.
- RAJA - No memory management, views are initialized with a raw pointer, index permutations are initialized at runtime, only dynamic dimensions.
- Kokkos - Configurable layout and default selection based on the memory/execution space, but only AoS and SoA are considered, even for
N > 2. For parallel work there is only one leading dimension - it does not map to 2D or 3D CUDA grids.
- CUV - Assumption that "everything is an n-dimensional array" (like Matlab), CPU and GPU support, column-major or row-major, integration with Python and Numpy.
Unstructured meshes
TNL has a unique data structure for the representation of unstructured meshes in high-performance simulation software, including GPU accelerators. See the original paper Configurable open-source data structure for distributed conforming unstructured homogeneous meshes with GPU support for the design choices an benchmarks comparing it to the MOAB library.
