George Zagaris
George Zagaris
Thanks @kennyweiss -- I am seeing this in just C++ code. I'll investigate further.
This sounds great!
The VTK mesh can be read into VisIt which then can dump the mesh in STL (binary and ASCII).
Hi @white238 -- generally I think I am OK with this. To clarify, the request is to have both ``, which would print the total number of ranks that generated...
> Don't we have that situation with some of our codes already? To the best of my knowledge, we're all building with `hdf5` and `conduit` which are the main external...
Note, along those lines, PR #1302 added a `RAJA_ENABLE_SIMD` CMake option
Thank you for the detailed explanation and suggestions @MrBurmark! The `RAJA::cuda_exec_occ_calc_async< BLOCK_SIZE >` sounds fancy! I'll definitely try it out. - Can this policy also be used for nested loops...
Thanks @MrBurmark -- it's always complicated with `RAJA::kernel` 😄 The example you provided is exactly what I was looking for!
> @gzagaris several years ago, we considered adding RAJA::initialize and RAJA::finalize methods. Kokkos does that, for example. We didn't really see a strong need at the time and thought RAJA...