Updates camp submodule - #2058
Conversation
|
|
||
| // _raja_res_wait_start | ||
| res_gpu2.wait_for(&e); | ||
| res_gpu2.wait_for(e); |
There was a problem hiding this comment.
This kind of change would be good to capture in the release note -- We typically wait until the end to update the release notes, but just marking this so we remember.
This is fine as the LoopData object is created on the stack in a function where the resource is an argument. Note that the LoopData object is used on the device in async kernels but that is also fine as the resource may not be used on the device.
|
@lc-hubcast approve |
|
To approve syncing this PR, please use the GitHub review comment feature to submit an approval. This ensures approval is tied to a specific commit to avoid the sync of malicious data. |
- Fixes a segfault
- More work is needed to better support SYCL's reduce and
be more similar to CUDA and HIP.
|
This looks good. We'll have to wait for this camp PR so the camp tests pass llnl/camp#213. |
In addition to the camp PR, both the |
|
@lc-hubcast approve |
|
To approve syncing this PR, please use the GitHub review comment feature to submit an approval. This ensures approval is tied to a specific commit to avoid the sync of malicious data. |
|
@smithsg84 @bechols97 is the idea to merge this, then fix the spack package issues upstream? I don't know what's the issue with the windows shared lib build. Apparently, I can't see the raw log file while other jobs are running. |
|
@rhornung67 That was my understanding. Also, I'm not sure on the Windows build either. It was passing prior to getting the branch up to date with develop. |
|
@bechols97 after some digging it seems that the windows build issue is due to using an experimental version of CMake (4.4), which has a known JSON string parsing issue tickled by changes to the camp array test. I'm going to push a change to your branch to a "stable" cmake version and see if that works. |
|
@bechols97 @smithsg84 my change to a stable version of CMake in our GHA Windows checks appears to have fixed the issue. |
There was a problem hiding this comment.
Does this need to be upstreamed to camp?
There was a problem hiding this comment.
@smithsg84 mentioned that this had been upstreamed in an earlier version of camp. This was just added to get the local spack packages working correctly.
6 participants
Summary
This PR updates the camp submodule to v2026.07.0. This includes build fixes for:
wait_foran event taking a reference instead of a pointerget_queuereturning a reference instead of a pointer