Andreas Borgen Longva
Andreas Borgen Longva
It seems that `gemm_tr` and `gemm_ad` currently do not leverage `matrixmultiply` for larger matrices. I noticed when I was profiling after making some changes to some of my performance-sensitive code....
#1017 introduced support for importing matrices stored in the matrix market format to `nalgebra-sparse`. We still need functionality for exporting matrices to this format.
The docs say: > If the matrix contains the sole reference to the sparsity pattern, then the data is returned as-is. Otherwise, the sparsity pattern is cloned. This stems from...
# Motivation A "matrix slice" is indeed analogous to Rust slices (`&[T]`), so the name isn't inherently problematic. However, the terminology quickly becomes ambiguous. For example, consider the following methods...
Currently, there are no guarantees about ordering of eigenvalues and singular vectors returned by nalgebra's decompositions. A common convention throughout mathematical literature is to assume that eigenvalues and singular values...
In #986 the `Debug` formatting of matrices was changed to forward to the storage. This works reasonably well for static matrices, but unfortunately it does not work well for dynamic...
We should probably write a separate README for `nalgebra-sparse` and make sure Documentation links and similar go to `docs.rs/nalgebra-sparse` instead of `docs.rs/nalgebra`.
In many cases, we might want to allocate a buffer that we fill with data. One way to do that might be to use an uninitialized buffer - but this...
Currently the import of matrices to the matrix market format (provided by #1017) is unable to import, for example, an integer matrix into a `CooMatrix`. For usability reasons it would...
This issue describes two in principle disjoint issues, but a design that address either needs to take care to accommodate both. They are: - Currently CSC/CSR (and COO) matrices only...