ElCapitan and Kokkos 5 and C++20#126
Merged
Merged
Conversation
adamdempsey90
commented
Apr 25, 2026
| set(Kokkos_ENABLE_SERIAL ON CACHE BOOL "" FORCE) | ||
| set(Kokkos_ENABLE_HIP_RELOCATABLE_DEVICE_CODE ON CACHE BOOL "" FORCE) | ||
| set(Kokkos_ENABLE_HIP_RELOCATABLE_DEVICE_CODE OFF CACHE BOOL "" FORCE) | ||
| set(Kokkos_ENABLE_IMPL_HIP_MALLOC_ASYNC OFF CACHE BOOL "" FORCE) |
Collaborator
Author
There was a problem hiding this comment.
This was needed to fix a segfault bug deep in kokkos
adamdempsey90
commented
Apr 25, 2026
| "${REBOUND_BUILD_DIR}/src/*.h") | ||
| add_custom_command(OUTPUT ${REBOUND_BUILD_DIR}/librebound.so | ||
| COMMAND make | ||
| COMMAND make CC=${CMAKE_C_COMPILER} |
Collaborator
Author
There was a problem hiding this comment.
Rebound has always just been using cc as the C compiler, but when using amdclang that was not being set propertly. We probably should've always been doing this though.
adamdempsey90
commented
Apr 25, 2026
Comment on lines
+148
to
+149
| "name": "elcap-gpu-debug", | ||
| "cacheVariables": { |
Collaborator
Author
There was a problem hiding this comment.
GPU debug builds are actually compiling very fast on these systems
Collaborator
Author
|
Looks like the disk tests are failing on all GPU systems (incl. cuda ones). Must be the update of either parthenon or eos |
pdmullen
approved these changes
May 19, 2026
adamdempsey90
commented
May 20, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Background
This updates the submodules to the latest versions and puts us on Kokkos 5 and C++20.
This was needed to get running on El Capitan systems. I've also refreshed / updated the modules/CMake stuff for ElCap.
This requires https://github.com/lanl/jaybenne/pulls
I'm still tracking down an issue where coming out of the pgen, the variables are being floored.That is now fixed.There is what appears to be a memory leak when using GPU aware MPI. Still investigating that.This was a missing argument toflux run.On the ElCap systems, you want to run in CPX mode, e.g. to get 2 nodes:
and then when running artemis you want to use 6 times the number of tasks as GPUs, so in the 2 node example you would do,
where 48 = 6 * 4 * 2 and
--exclusivemakes sure the task->gpu bindings are correct.Running the binary problem with AMR in 2D cylindrical looks good.
Description of Changes
Checklist
// This file was created in part or in whole by generative AI