Vijay M
Thanks for the response. The input vector for which I need to compute SpMV operations is essentially a 2-D dense matrix that is provided by a Fortran subroutine. And the...
> What I can't follow though is how the performance of A*x can profit from vectorization when x is stored in column-major? In my understanding, this would mean that entries...
For the current application use-case of interest, the row-length `N` is typically above 100K, and col-length `n` is typically between 10-40. At the moment, we are looking at optimizations of...
## Review checklist for @vijaysm ### Conflict of interest - [x] I confirm that I have read the [JOSS conflict of interest (COI) policy](https://github.com/openjournals/joss/blob/master/COI.md) and that: I have no COIs...
The paper is well written. But I think it can be improved by fixing several spelling mistakes in the text and some missing punctuation. Some specific comments: - P2. L48:...
I am happy with the responses of the author to address all the questions I've raised. The software and the manuscript are in good shape now as well. I recommend...
Thanks @yangbai90
I'd like to second this comment as a reviewer. There are a lot of very successful and tested, general-purpose FEM libraries out there. I fail to see a proper motivation...
The manuscript reads much better now in the statement of need section.
While installing PETSc is less of a hassle in most systems, having a docker image will allow interested users to get started quickly. So I agree that this should be...