Arturo Vargas

Warnings with xlC compiler

3

Hi all, I recently built AMGX using xlC (Version: 16) as my host compiler, while building the code I hit a number of warnings: ``warning: 1540-5200 The option "-dynamic" is...

Setting HIP compiler

3

Hello Team Hypre, I was wondering how one can manually set the hip compiler. Looking at the configure file: https://github.com/hypre-space/hypre/blob/514c72be691b8698df7e7069d54458831cbc531e/src/config/configure.in#L1574 It looks like if we use the variable for the...

How to cite CHAI?

1

Hello! Is there a preferred way to cite CHAI?

GPU Support

4

Hi, I was wondering if GPU support (Nvidia/AMD) is on the roadmap?

EA data layout vs .GetSubMatrix(ndofs, ndofs, A_loc) layout

1

Although not directly exposed to the user, I noticed that EA data is layout in a row major format vs .GetSubMatrix() which returns a submatrix in column major. It would...

Add ability to execute dynamic forall based on a string name policy

2

See discussion: https://github.com/LLNL/RAJA/pull/1280#discussion_r976857762

A potential way to do this is to introduce a multiloop method to handle perfectly nested loops. This could be a gateway for loop reordering as well.

Update RAJA for v2023.06.00 and beyond

2

RAJA loop_exec will be removed, seq_exec will be taking its place. | PR | Author | Editor | Reviewers | Assignment | Approval | Merge| | --- | --- |...

Wrappers for math lib batch calls

14

With some macro magic I learned from @tomstitt and some existing work from @pazner we have started creating small wrappers for CUDA/HIP library calls to do batch matrix operations. Here...