-
Notifications
You must be signed in to change notification settings - Fork 53
Develop stream 2024-11-07 #486
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Develop stream 2024-11-07 #486
Conversation
64d84b9
to
a0e0108
Compare
Where optional<T> inherits optional<T>::construct via a series of classes, optional<T&> does not. This means that optional<T&>::emplace() was broken and called into a member function that did not exist. This replaces the functionality to make optional<T&>::emplace() change the stored reference to the new one. Note that it does _not_ emplace the referee, as this would lead to questionable behavior when the optional holds nullopt. This was revealed by a change in LLVM, see llvm/llvm-project#90152 and ROCm#404.
… is supported from cuda::std
add large_data only to copy tests
Co-authored-by: Milo Lurati <[email protected]>
a0e0108
to
0a39100
Compare
I rebased this PR. Looking at the commits this will need the PR in rocprim, ROCm/rocPRIM#631. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm, all unit tests for this pr still passing on gfx908 and gfx1101.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Tests look good for now (although the failures in the unspecified devops CI system are worrisome). I noticed the new changelog entry should be for 6.4. We can update that ourselves in a separate PR after merging this one in.
This PR also includes commits from this PR which should be merged before. The new commits that are included add: