|
Ensign
Low-rank framework for multi-core CPUs and GPUs
|
| docs | |
| ▼ include | |
| ▼ generic | |
| common.hpp | Provides shared types, platform headers, and basic definitions |
| fft.hpp | Applies batched FFTs over selected dimensions, commonly one per rank index |
| index.hpp | Converts between vector indices and combined linear indices |
| kernels.hpp | Declares CUDA kernels for elementwise operations and numerical updates |
| matrix.hpp | Declares dense matrix and vector operations, including BLAS wrappers |
| netcdf.hpp | Provides a NetCDF writer for simulation output |
| storage.hpp | Provides contiguous multidimensional arrays for CPU and GPU storage |
| tensor.hpp | Provides tensor matricization, tensorization, and orthogonalization |
| timer.hpp | Provides elapsed-time measurement and named timers |
| tree.hpp | Provides node types for binary tree tensor networks |
| utility.hpp | Provides numerical helpers and utilities for arrays and indices |
| ▼ lr | |
| coefficients.hpp | Computes weighted coefficients and integrals of low-rank factors (the c and d factors in the notation of https://doi.org/10.1137/18M116383X) |
| interpolatory.hpp | Provides interpolation-based construction of a low-rank representations |
| lr.hpp | Provides two-factor (i.e. matrix) low-rank types and their operations |