[Vectorizer] Add vectorization support for fixed smul/umul intrinsics

This requires a couple of tweaks to existing vectorization functions as they were assuming that only the second call argument (ctlz/cttz/powi) could ever be the 'always scalar' argument, but for smul.fix + umul.fix its the third argument.

Differential Revision: https://reviews.llvm.org/D58616

llvm-svn: 354790

llvm/lib/Analysis/VectorUtils.cpp

@@ -52,6 +52,8 @@ bool llvm::isTriviallyVectorizable(Intrinsic::ID ID) {
   case Intrinsic::ssub_sat:
   case Intrinsic::uadd_sat:
   case Intrinsic::usub_sat:
+  case Intrinsic::smul_fix:
+  case Intrinsic::umul_fix:
   case Intrinsic::sqrt: // Begin floating-point.
   case Intrinsic::sin:
   case Intrinsic::cos:
@@ -92,6 +94,9 @@ bool llvm::hasVectorInstrinsicScalarOpd(Intrinsic::ID ID,
   case Intrinsic::cttz:
   case Intrinsic::powi:
     return (ScalarOpdIdx == 1);
+  case Intrinsic::smul_fix:
+  case Intrinsic::umul_fix:
+    return (ScalarOpdIdx == 2);
   default:
     return false;
   }

llvm/lib/Transforms/Vectorize/LoopVectorizationLegality.cpp

@@ -713,18 +713,21 @@ bool LoopVectorizationLegality::canVectorizeInstrs() {
         return false;
       }
 
-      // Intrinsics such as powi,cttz and ctlz are legal to vectorize if the
-      // second argument is the same (i.e. loop invariant)
-      if (CI && hasVectorInstrinsicScalarOpd(
-                    getVectorIntrinsicIDForCall(CI, TLI), 1)) {
+      // Some intrinsics have scalar arguments and should be same in order for
+      // them to be vectorized (i.e. loop invariant).
+      if (CI) {
         auto *SE = PSE.getSE();
-        if (!SE->isLoopInvariant(PSE.getSCEV(CI->getOperand(1)), TheLoop)) {
-          ORE->emit(createMissedAnalysis("CantVectorizeIntrinsic", CI)
-                    << "intrinsic instruction cannot be vectorized");
-          LLVM_DEBUG(dbgs()
-                     << "LV: Found unvectorizable intrinsic " << *CI << "\n");
-          return false;
-        }
+        Intrinsic::ID IntrinID = getVectorIntrinsicIDForCall(CI, TLI);
+        for (unsigned i = 0, e = CI->getNumArgOperands(); i != e; ++i)
+          if (hasVectorInstrinsicScalarOpd(IntrinID, i)) {
+            if (!SE->isLoopInvariant(PSE.getSCEV(CI->getOperand(i)), TheLoop)) {
+              ORE->emit(createMissedAnalysis("CantVectorizeIntrinsic", CI)
+                        << "intrinsic instruction cannot be vectorized");
+              LLVM_DEBUG(dbgs() << "LV: Found unvectorizable intrinsic " << *CI
+                                << "\n");
+              return false;
+            }
+          }
       }
 
       // Check that the instruction return type is vectorizable.

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

@@ -437,8 +437,9 @@ static bool InTreeUserNeedToExtract(Value *Scalar, Instruction *UserInst,
   case Instruction::Call: {
     CallInst *CI = cast<CallInst>(UserInst);
     Intrinsic::ID ID = getVectorIntrinsicIDForCall(CI, TLI);
-    if (hasVectorInstrinsicScalarOpd(ID, 1)) {
-      return (CI->getArgOperand(1) == Scalar);
+    for (unsigned i = 0, e = CI->getNumArgOperands(); i != e; ++i) {
+      if (hasVectorInstrinsicScalarOpd(ID, i))
+        return (CI->getArgOperand(i) == Scalar);
     }
     LLVM_FALLTHROUGH;
   }
@@ -1860,9 +1861,11 @@ void BoUpSLP::buildTree_rec(ArrayRef<Value *> VL, unsigned Depth,
         return;
       }
       Function *Int = CI->getCalledFunction();
-      Value *A1I = nullptr;
-      if (hasVectorInstrinsicScalarOpd(ID, 1))
-        A1I = CI->getArgOperand(1);
+      unsigned NumArgs = CI->getNumArgOperands();
+      SmallVector<Value*, 4> ScalarArgs(NumArgs, nullptr);
+      for (unsigned j = 0; j != NumArgs; ++j)
+        if (hasVectorInstrinsicScalarOpd(ID, j))
+          ScalarArgs[j] = CI->getArgOperand(j);
       for (unsigned i = 1, e = VL.size(); i != e; ++i) {
         CallInst *CI2 = dyn_cast<CallInst>(VL[i]);
         if (!CI2 || CI2->getCalledFunction() != Int ||
@@ -1874,16 +1877,19 @@ void BoUpSLP::buildTree_rec(ArrayRef<Value *> VL, unsigned Depth,
                             << "\n");
           return;
         }
-        // ctlz,cttz and powi are special intrinsics whose second argument
-        // should be same in order for them to be vectorized.
-        if (hasVectorInstrinsicScalarOpd(ID, 1)) {
-          Value *A1J = CI2->getArgOperand(1);
-          if (A1I != A1J) {
-            BS.cancelScheduling(VL, VL0);
-            newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
-            LLVM_DEBUG(dbgs() << "SLP: mismatched arguments in call:" << *CI
-                              << " argument " << A1I << "!=" << A1J << "\n");
-            return;
+        // Some intrinsics have scalar arguments and should be same in order for
+        // them to be vectorized.
+        for (unsigned j = 0; j != NumArgs; ++j) {
+          if (hasVectorInstrinsicScalarOpd(ID, j)) {
+            Value *A1J = CI2->getArgOperand(j);
+            if (ScalarArgs[j] != A1J) {
+              BS.cancelScheduling(VL, VL0);
+              newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+              LLVM_DEBUG(dbgs() << "SLP: mismatched arguments in call:" << *CI
+                                << " argument " << ScalarArgs[j] << "!=" << A1J
+                                << "\n");
+              return;
+            }
           }
         }
         // Verify that the bundle operands are identical between the two calls.
@@ -3443,9 +3449,9 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) {
       std::vector<Value *> OpVecs;
       for (int j = 0, e = CI->getNumArgOperands(); j < e; ++j) {
         ValueList OpVL;
-        // ctlz,cttz and powi are special intrinsics whose second argument is
-        // a scalar. This argument should not be vectorized.
-        if (hasVectorInstrinsicScalarOpd(IID, 1) && j == 1) {
+        // Some intrinsics have scalar arguments. This argument should not be
+        // vectorized.
+        if (hasVectorInstrinsicScalarOpd(IID, j)) {
           CallInst *CEI = cast<CallInst>(VL0);
           ScalarArg = CEI->getArgOperand(j);
           OpVecs.push_back(CEI->getArgOperand(j));