[PATCH] D93707: [FPEnv] Allow fneg + strict_fsub -> strict_fadd in DAGCombiner
Craig Topper via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Dec 22 09:42:34 PST 2020
craig.topper added inline comments.
================
Comment at: llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp:13416
+ return DAG.getNode(ISD::STRICT_FADD, DL, DAG.getVTList(VT, ChainVT),
+ {Chain, N0, NegN1}, Flags);
+
----------------
Can we use SelectionDAG::FlagInserter like visitFSUB instead of passing Flags?
================
Comment at: llvm/test/CodeGen/X86/fp-intrinsics.ll:70
; X87-NEXT: fldz
-; X87-NEXT: fsubrl {{[0-9]+}}(%esp)
+; X87-NEXT: fchs
+; X87-NEXT: faddl 4(%esp)
----------------
Why did this get worse?
================
Comment at: llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll:680
define <1 x float> @constrained_vector_fsub_v1f32() #0 {
+; CHECK: # float -1
; CHECK-LABEL: constrained_vector_fsub_v1f32:
----------------
The checks line are normally generated by a script update_llc_test_checks.py. The next time someone reruns that script on this file, theses lines will disappear.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D93707/new/
https://reviews.llvm.org/D93707
More information about the llvm-commits
mailing list