[PATCH] D150768: [SVE ACLE] Canonicalise SVE merging intrinsics

[Jolanta Jensen via Phabricator via llvm-commits]

jolanta.jensen created this revision.
jolanta.jensen added reviewers: mgabka, paulwalker-arm.
Herald added subscribers: ctetreau, hiraditya, tschuett.
Herald added a project: All.
jolanta.jensen requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

Canonicalise SVE merging intrinsics to their equivalent
undef (_u) variants when they take an all active predicate.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D150768

Files:
  llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp


Index: llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
===================================================================
--- llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -1265,11 +1265,21 @@
 
 static std::optional<Instruction *> instCombineSVEVectorAdd(InstCombiner &IC,
                                                             IntrinsicInst &II) {
+  if (auto FMLA_U =
+          instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul_u,
+                                            Intrinsic::aarch64_sve_fmla_u>(
+              IC, II, true))
+    return FMLA_U;
   if (auto FMLA =
           instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul,
                                             Intrinsic::aarch64_sve_fmla>(IC, II,
                                                                          true))
     return FMLA;
+  if (auto MLA_U =
+          instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul_u,
+                                            Intrinsic::aarch64_sve_mla_u>(
+              IC, II, true))
+    return MLA_U;
   if (auto MLA = instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul,
                                                    Intrinsic::aarch64_sve_mla>(
           IC, II, true))
@@ -1288,11 +1298,21 @@
 
 static std::optional<Instruction *> instCombineSVEVectorSub(InstCombiner &IC,
                                                             IntrinsicInst &II) {
+  if (auto FMLS_U =
+          instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul_u,
+                                            Intrinsic::aarch64_sve_fmls_u>(
+              IC, II, true))
+    return FMLS_U;
   if (auto FMLS =
           instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul,
                                             Intrinsic::aarch64_sve_fmls>(IC, II,
                                                                          true))
     return FMLS;
+  if (auto MLS_U =
+          instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul_u,
+                                            Intrinsic::aarch64_sve_mls_u>(
+              IC, II, true))
+    return MLS_U;
   if (auto MLS = instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul,
                                                    Intrinsic::aarch64_sve_mls>(
           IC, II, true))
@@ -1677,27 +1697,15 @@
   case Intrinsic::aarch64_sve_fmul:
     return instCombineSVEVectorMul(IC, II);
   case Intrinsic::aarch64_sve_fadd:
-  case Intrinsic::aarch64_sve_add:
-    return instCombineSVEVectorAdd(IC, II);
   case Intrinsic::aarch64_sve_fadd_u:
-    return instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul_u,
-                                             Intrinsic::aarch64_sve_fmla_u>(
-        IC, II, true);
+  case Intrinsic::aarch64_sve_add:
   case Intrinsic::aarch64_sve_add_u:
-    return instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul_u,
-                                             Intrinsic::aarch64_sve_mla_u>(
-        IC, II, true);
+    return instCombineSVEVectorAdd(IC, II);
   case Intrinsic::aarch64_sve_fsub:
-  case Intrinsic::aarch64_sve_sub:
-    return instCombineSVEVectorSub(IC, II);
   case Intrinsic::aarch64_sve_fsub_u:
-    return instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_fmul_u,
-                                             Intrinsic::aarch64_sve_fmls_u>(
-        IC, II, true);
+  case Intrinsic::aarch64_sve_sub:
   case Intrinsic::aarch64_sve_sub_u:
-    return instCombineSVEVectorFuseMulAddSub<Intrinsic::aarch64_sve_mul_u,
-                                             Intrinsic::aarch64_sve_mls_u>(
-        IC, II, true);
+    return instCombineSVEVectorSub(IC, II);
   case Intrinsic::aarch64_sve_tbl:
     return instCombineSVETBL(IC, II);
   case Intrinsic::aarch64_sve_uunpkhi:


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D150768.522991.patch
Type: text/x-patch
Size: 3917 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230517/cb1006d2/attachment.bin>