From 3b416db19ba91207649430cfa4604ffd331bf86c Mon Sep 17 00:00:00 2001
From: Farhana Aleen <farhana.aleen@gmail.com>
Date: Mon, 2 Jul 2018 17:55:31 +0000
Subject: [SLP] Recognize min/max pattern using instructions producing same
 values.

Summary: It is common to have the following min/max pattern during the intermediate stages of SLP since we only optimize at the end. This patch tries to catch such patterns and allow more vectorization.

         %1 = extractelement <2 x i32> %a, i32 0
         %2 = extractelement <2 x i32> %a, i32 1
         %cond = icmp sgt i32 %1, %2
         %3 = extractelement <2 x i32> %a, i32 0
         %4 = extractelement <2 x i32> %a, i32 1
         %select = select i1 %cond, i32 %3, i32 %4

Author: FarhanaAleen

Reviewed By: ABataev, RKSimon, spatel

Differential Revision: https://reviews.llvm.org/D47608

llvm-svn: 336130
---
 llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp | 71 +++++++++++++++++++++++++
 1 file changed, 71 insertions(+)

(limited to 'llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp')

diff --git a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
index 81878b7abb1..dd654b7b493 100644
--- a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
+++ b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
@@ -5469,6 +5469,77 @@ class HorizontalReduction {
         return OperationData(
             Instruction::FCmp, LHS, RHS, RK_Max,
             cast<Instruction>(Select->getCondition())->hasNoNaNs());
+      } else {
+        // Try harder: look for min/max pattern based on instructions producing
+        // same values such as: select ((cmp Inst1, Inst2), Inst1, Inst2).
+        // During the intermediate stages of SLP, it's very common to have
+        // pattern like this (since optimizeGatherSequence is run only once
+        // at the end):
+        // %1 = extractelement <2 x i32> %a, i32 0
+        // %2 = extractelement <2 x i32> %a, i32 1
+        // %cond = icmp sgt i32 %1, %2
+        // %3 = extractelement <2 x i32> %a, i32 0
+        // %4 = extractelement <2 x i32> %a, i32 1
+        // %select = select i1 %cond, i32 %3, i32 %4
+        CmpInst::Predicate Pred;
+        Instruction *L1;
+        Instruction *L2;
+
+        LHS = Select->getTrueValue();
+        RHS = Select->getFalseValue();
+        Value *Cond = Select->getCondition();
+
+        // TODO: Support inverse predicates.
+        if (match(Cond, m_Cmp(Pred, m_Specific(LHS), m_Instruction(L2)))) {
+          if (!isa<ExtractElementInst>(RHS) ||
+              !L2->isIdenticalTo(cast<Instruction>(RHS)))
+            return OperationData(V);
+        } else if (match(Cond, m_Cmp(Pred, m_Instruction(L1), m_Specific(RHS)))) {
+          if (!isa<ExtractElementInst>(LHS) ||
+              !L1->isIdenticalTo(cast<Instruction>(LHS)))
+            return OperationData(V);
+        } else {
+          if (!isa<ExtractElementInst>(LHS) || !isa<ExtractElementInst>(RHS))
+            return OperationData(V);
+          if (!match(Cond, m_Cmp(Pred, m_Instruction(L1), m_Instruction(L2))) ||
+              !L1->isIdenticalTo(cast<Instruction>(LHS)) ||
+              !L2->isIdenticalTo(cast<Instruction>(RHS)))
+            return OperationData(V);
+        }
+        switch (Pred) {
+        default:
+          return OperationData(V);
+
+        case CmpInst::ICMP_ULT:
+        case CmpInst::ICMP_ULE:
+          return OperationData(Instruction::ICmp, LHS, RHS, RK_UMin);
+
+        case CmpInst::ICMP_SLT:
+        case CmpInst::ICMP_SLE:
+          return OperationData(Instruction::ICmp, LHS, RHS, RK_Min);
+
+        case CmpInst::FCMP_OLT:
+        case CmpInst::FCMP_OLE:
+        case CmpInst::FCMP_ULT:
+        case CmpInst::FCMP_ULE:
+          return OperationData(Instruction::FCmp, LHS, RHS, RK_Min,
+                               cast<Instruction>(Cond)->hasNoNaNs());
+
+        case CmpInst::ICMP_UGT:
+        case CmpInst::ICMP_UGE:
+          return OperationData(Instruction::ICmp, LHS, RHS, RK_UMax);
+
+        case CmpInst::ICMP_SGT:
+        case CmpInst::ICMP_SGE:
+          return OperationData(Instruction::ICmp, LHS, RHS, RK_Max);
+
+        case CmpInst::FCMP_OGT:
+        case CmpInst::FCMP_OGE:
+        case CmpInst::FCMP_UGT:
+        case CmpInst::FCMP_UGE:
+          return OperationData(Instruction::FCmp, LHS, RHS, RK_Max,
+                               cast<Instruction>(Cond)->hasNoNaNs());
+        }
       }
     }
     return OperationData(V);
-- 
cgit v1.2.3