[PATCH] D110177: [AArch64][SVE] Remove redundant PTEST following PNEXT/PFIRST

[Peter Waller via Phabricator via llvm-commits]

peterwaller-arm added inline comments.


================
Comment at: llvm/test/CodeGen/AArch64/sve-ptest-removal-pfirst.mir:3
+
+# Test instruction sequences where PTEST is redundant and thus gets removed.
+---
----------------
sdesmalen wrote:
> Do these tests necessarily need to be MIR tests? The reason I ask is because I generally find them more difficult to read and would have thought some LLVM IR intrinsics would have been sufficient to prove an extra `ptest` is not added.
Good question. These tests reuse the same form as the other tests for this. But I agree an IR test would be seem straightforward. I'll wait for anyone else to chime in with reasoning and if I don't hear anything I'll switch it on Monday next week. I have an open question in my mind of whether the other tests (llvm/test/CodeGen/AArch64/sve-ptest-removal-*.mir) should be migrated to IR tests too, as a separate commit.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110177/new/

https://reviews.llvm.org/D110177