[PATCH] D83395: [SVE] Code generation for fixed length vector truncates.
Eli Friedman via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Wed Jul 8 12:45:56 PDT 2020
efriedma added a comment.
Probably worth adding a testcase for truncating from `<4 x i64>` to `<4 x i8>`.
================
Comment at: llvm/lib/Target/AArch64/AArch64ISelLowering.cpp:1063
setOperationAction(ISD::STORE, VT, Custom);
+ setOperationAction(ISD::TRUNCATE, VT, Custom);
}
----------------
This specifically applies to the result type. You might want to note that you're implicitly depending on the fact that we do custom legalization for NEON TRUNCATE operations for other reasons.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D83395/new/
https://reviews.llvm.org/D83395
More information about the llvm-commits
mailing list