fritzgoebel
fritzgoebel
This PR adds a parameter to the triangular solvers for choosing the algorithm to be used rather than making this depend on the matrix spmv strategy. Currently it is only...
This PR adds an interface for wrapping an inner operator (like a direct or iterative solver) into a (for now symmetrical, e.g. RCM) reordering via a reordering factory and diagonal...
This PR adds mixed precision support to the naive caching triangular solvers via `mixed_precision_dispatch_real_complex`. Also adds the kernel not relying on independent thread scheduling to hip.
This PR adds an example enabling users to perform simple strong and weak scaling studies for the distributed SpMV. The example can be run with 2D (5-pt or 9-pt) and...
This PR makes the ISAI preconditioner support reduced precision storage format. It is based on Tobias' mp spmv approach for ell, which is slightly changed to use an accessor on...
Add MC64
This PR adds the MC64 reordering and equilibration algorithm. MC64 moves large matrix entries to the diagonal, this implementation supports two strategies: maximizing the sum or the product of the...
This PR should fix issues #540 and #542. It also removes the nonsymmetric ginkgo solver as previously suggested by @pelesh. The problem were a missing host side copy of the...
This PR updates the Ginkgo interface in order to be able to handle data coming from the GPU. Currently this means: - in the first solver call: Copying matrices to...
When trying to build hiop on Frontier with the build script from the exago kpp2 branch (https://github.com/pnnl/ExaGO/tree/kpp2-branch) I get a lot of warnings as errors. 
In `src/Linalg/LinAlgFactory.cpp`, in line 366, a `hiop::hiopMatrixDenseRaja` is created outside of `HIOP_USE_RAJA`. This gives an error when building without Raja.