The program will be continually updated as we progress with the organization of the workshop.
The final program should be online shortly after the deadline for registering contributed talks has passed on July 1st, 2025.
Like new features planned for GPUs, for example for output writing, state of some packages like KOKKOS. Or this could eventually be covered for example in a workshop that focuses on reproducibility and good/bad practices when working with GPU accelerator packages.
Which features are available for GPUs can be seen on the command and style overview tables as character codes in parenthesis (e.g. here: 5.5. General commands — LAMMPS documentation) . New ports are added incrementally and rarely is there a large bunch of them that warrants a special announcement. Support for new hardware is rare and general changes in the GPU acceleration approach even more so. For KOKKOS, support is dependent on the support in the Kokkos library itself (but even if Kokkos supports it, some adjustments may be needed, e.g. when Kokkos support for a platform is not complete, or a platform does not yet have a feature for that platform that is used in LAMMPS, or LAMMPS has specializations for different Kokkos platforms).
You need to explain, how GPUs are be used for writing output.
If you want to see that, you need to gather a team and organize such a workshop.
We covered porting LAMMPS features to KOKKOS at a workshop last year. Stan’s lectures were recorded and are available online on YouTube
I normally give a “State of the KOKKOS Package” talk at the LAMMPS workshop, which would cover recent our work. Planning to do that again, pending enough space in the schedule.