aboutsummaryrefslogtreecommitdiff
path: root/llvm/docs/tutorial/MyFirstLanguageFrontend
diff options
context:
space:
mode:
authorJoseph Huber <jhuber6@vols.utk.edu>2023-05-24 07:59:37 -0500
committerJoseph Huber <jhuber6@vols.utk.edu>2023-05-31 17:12:43 -0500
commit349c0aacb38072a868421ce7c460514be57a3de7 (patch)
tree6031f34800bdd1ce43668cff3a297fb89a3d8c2a /llvm/docs/tutorial/MyFirstLanguageFrontend
parent54e47724cfbf12dce05412c5b127ae311066afb3 (diff)
downloadllvm-349c0aacb38072a868421ce7c460514be57a3de7.zip
llvm-349c0aacb38072a868421ce7c460514be57a3de7.tar.gz
llvm-349c0aacb38072a868421ce7c460514be57a3de7.tar.bz2
[OpenMP] Remove 'keep_alive' functionality from the device RTL
The OpenMP DeviceRTL uses a hacky workaround to keep certain runtime calls alive. This used a function that prevented them from being optimized out. We needed this hack because the 'OpenMPOpt' pass likes to introduce new runtime calls into the TU. This then interacted badly with the method of linking the bitcode file per-TU like we do with Nvidia. The OpenMPOpt pass would then generate a runtime call to a function that was never linked in. This should not be a problem anymore because we unconditionally link in the `libomptarget.devicertl.a` runtime library. This should thus only extract symbols that are undefined. So, if we do end up with an unresolved reference it will be resolved by the static library. The downside to this is that if we are doing non-LTO NVPTX compilation that introduces one of these calls it will be linked outside the module and therefore provide the overhead of an external function call. However, removing this flag should make optimizing things easier. We will need to see if that performance is a problem. Reviewed By: ye-luo Differential Revision: https://reviews.llvm.org/D151324
Diffstat (limited to 'llvm/docs/tutorial/MyFirstLanguageFrontend')
0 files changed, 0 insertions, 0 deletions