Luc Berger
Luc Berger
Hum, that leads to a non-symmetric graph: ``` [0 1 1] [0 0 0] [0 0 0] ``` you could switch to the following graph: ``` [0 1 1] [1...
I am not sure what your questions is? We have a TPL specialization layer that defines a struct with a `value` field set to `true` or `false` depending on the...
Should we be less aggressive and try to rebuild the documentation only once a day?
PR #804 fixes the build issue but does not provide an implementation for SpGEMM cusparse backend with cuda/11. Providing the backend for cuda/11 or later will require to rewrite the...
This makes sense although I would like to think about all the TPLs together and try to come up with an implementation that takes care of all of this for...
Correct, I think handling the TPLs can be done in two parts 1. augment the CrsMatrix to hold additional data generated by the TPL 2. keep the individual kernel implementations...
In general we strive to have Kokkos Core and Kokkos Kernels both work with the head of their respective develop. To that end we have eliminated dependencies in Kokkos Kernels...
Looking at the constructor body, it seems that we are assuming that the user will give us both `row_map` and `colindices` as `OrdinalType*` so this is not a bug but...
@brian-kelley we could do that or just add an overload, but the question really becomes: do we want to keep the current version or should we just deprecate it? Do...
Hum, I have mixed feelings about this, I see how this is convenient for test and examples but I suspect that if we allow this actual applications will use it...