Bump the LLVM version to the latest official LLVM 16 release #633
Labels
build
The issue related to the CUDA Quantum build
Milestone
Comments
|
Latest is 16.0.6. https://github.com/llvm/llvm-project/releases/tag/llvmorg-16.0.6 |
|
LLVM 17 drops typed pointer support. |
schweitzpgi
added a commit
to schweitzpgi/cuda-quantum
that referenced
this issue
Sep 20, 2023
This changes the internal llvm (tpls/llvm) to point to a SHA on the release/16.x branch. This SHA corresponds to version 16.0.6. Documentation should be updated to refer to LLVM 16.0.6 as the latest version supported.
schweitzpgi
added a commit
to schweitzpgi/cuda-quantum
that referenced
this issue
Sep 20, 2023
This changes the internal llvm (tpls/llvm) to point to a SHA on the release/16.x branch. This SHA corresponds to version 16.0.6. Documentation should be updated to refer to LLVM 16.0.6 as the latest version supported. Add workaround for unsigned integer constants as template arguments.
schweitzpgi
added a commit
to schweitzpgi/cuda-quantum
that referenced
this issue
Sep 20, 2023
This changes the internal llvm (tpls/llvm) to point to a SHA on the release/16.x branch. This SHA corresponds to version 16.0.6. Documentation should be updated to refer to LLVM 16.0.6 as the latest version supported. Add workaround for unsigned integer constants as template arguments. Move the subdirectory up in the file to avoid it getting "lost". It's not clear what the bug is towards the end of the file that was causing it to be omitted in my workspace setup.
MarkusPfundstein
pushed a commit
to fermioniq/cuda-quantum
that referenced
this issue
Sep 23, 2024
* Fix NVIDIA#633: Bump the LLVM version to 16.0.6. This changes the internal llvm (tpls/llvm) to point to a SHA on the release/16.x branch. This SHA corresponds to version 16.0.6. Documentation should be updated to refer to LLVM 16.0.6 as the latest version supported. Add workaround for unsigned integer constants as template arguments. Move the subdirectory up in the file to avoid it getting "lost". It's not clear what the bug is towards the end of the file that was causing it to be omitted in my workspace setup. * Update lib/Frontend/nvqpp/ConvertDecl.cpp Co-authored-by: Ben Howe <[email protected]> --------- Co-authored-by: Ben Howe <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
See issue #16 as to why the project is stuck on LLVM 16.
We can and should try bumping
tpls/llvmto point to the latest version of LLVM on the release 16 branch. If that is copacetic, then make that the version of LLVM used by the CI, shipped in the container images, etc.As it stands now,
tpls/llvmis stuck on an arbitrary commit before release 16 (pre-16).See also issue #609 .
The text was updated successfully, but these errors were encountered: