[PATCH] D101582: [InstCombine] cttz(zext(x)) -> zext(cttz(x)) if the 'ZeroIsUndef' parameter is 'true' (PR50172)

[Dávid Bolvanský via Phabricator via llvm-commits]

xbolva00 updated this revision to Diff 341680.
xbolva00 added a comment.

Added tests


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D101582/new/

https://reviews.llvm.org/D101582

Files:
  llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
  llvm/test/Transforms/InstCombine/cttz.ll


Index: llvm/test/Transforms/InstCombine/cttz.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/InstCombine/cttz.ll
@@ -0,0 +1,63 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -S -instcombine | FileCheck %s
+
+declare i32 @llvm.cttz.i32(i32, i1)
+declare <2 x i64> @llvm.cttz.v2i64(<2 x i64>, i1)
+declare void @use(i32)
+
+define i32 @cttz_zext_zero_undef(i16 %x) {
+; CHECK-LABEL: @cttz_zext_zero_undef(
+; CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.cttz.i16(i16 [[X:%.*]], i1 true), !range [[RNG0:![0-9]+]]
+; CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[TMP1]] to i32
+; CHECK-NEXT:    ret i32 [[TMP2]]
+;
+  %z = zext i16 %x to i32
+  %tz = call i32 @llvm.cttz.i32(i32 %z, i1 true)
+  ret i32 %tz
+}
+
+define i32 @cttz_zext_zero_def(i16 %x) {
+; CHECK-LABEL: @cttz_zext_zero_def(
+; CHECK-NEXT:    [[Z:%.*]] = zext i16 [[X:%.*]] to i32
+; CHECK-NEXT:    [[TZ:%.*]] = call i32 @llvm.cttz.i32(i32 [[Z]], i1 false), !range [[RNG1:![0-9]+]]
+; CHECK-NEXT:    ret i32 [[TZ]]
+;
+  %z = zext i16 %x to i32
+  %tz = call i32 @llvm.cttz.i32(i32 %z, i1 false)
+  ret i32 %tz
+}
+
+define i32 @cttz_zext_zero_undef_extra_use(i16 %x) {
+; CHECK-LABEL: @cttz_zext_zero_undef_extra_use(
+; CHECK-NEXT:    [[Z:%.*]] = zext i16 [[X:%.*]] to i32
+; CHECK-NEXT:    call void @use(i32 [[Z]])
+; CHECK-NEXT:    [[TZ:%.*]] = call i32 @llvm.cttz.i32(i32 [[Z]], i1 true), !range [[RNG1]]
+; CHECK-NEXT:    ret i32 [[TZ]]
+;
+  %z = zext i16 %x to i32
+  call void @use(i32 %z)
+  %tz = call i32 @llvm.cttz.i32(i32 %z, i1 true)
+  ret i32 %tz
+}
+
+define <2 x i64> @cttz_zext_zero_undef_vec(<2 x i32> %x) {
+; CHECK-LABEL: @cttz_zext_zero_undef_vec(
+; CHECK-NEXT:    [[TMP1:%.*]] = call <2 x i32> @llvm.cttz.v2i32(<2 x i32> [[X:%.*]], i1 true)
+; CHECK-NEXT:    [[TMP2:%.*]] = zext <2 x i32> [[TMP1]] to <2 x i64>
+; CHECK-NEXT:    ret <2 x i64> [[TMP2]]
+;
+  %z = zext <2 x i32> %x to <2 x i64>
+  %tz = tail call <2 x i64> @llvm.cttz.v2i64(<2 x i64> %z, i1 true)
+  ret <2 x i64> %tz
+}
+
+define <2 x i64> @cttz_zext_zero_def_vec(<2 x i32> %x) {
+; CHECK-LABEL: @cttz_zext_zero_def_vec(
+; CHECK-NEXT:    [[Z:%.*]] = zext <2 x i32> [[X:%.*]] to <2 x i64>
+; CHECK-NEXT:    [[TZ:%.*]] = tail call <2 x i64> @llvm.cttz.v2i64(<2 x i64> [[Z]], i1 false)
+; CHECK-NEXT:    ret <2 x i64> [[TZ]]
+;
+  %z = zext <2 x i32> %x to <2 x i64>
+  %tz = tail call <2 x i64> @llvm.cttz.v2i64(<2 x i64> %z, i1 false)
+  ret <2 x i64> %tz
+}
Index: llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
===================================================================
--- llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
+++ llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
@@ -424,6 +424,7 @@
          "Expected cttz or ctlz intrinsic");
   bool IsTZ = II.getIntrinsicID() == Intrinsic::cttz;
   Value *Op0 = II.getArgOperand(0);
+  Value *Op1 = II.getArgOperand(1);
   Value *X;
   // ctlz(bitreverse(x)) -> cttz(x)
   // cttz(bitreverse(x)) -> ctlz(x)
@@ -438,6 +439,15 @@
     if (match(Op0, m_Neg(m_Value(X))))
       return IC.replaceOperand(II, 0, X);
 
+    // Zext doesn't change the number of trailing zeros, so narrow:
+    // cttz(zext(x)) -> zext(cttz(x)) if the 'ZeroIsUndef' parameter is 'true'.
+    if (match(Op0, m_OneUse(m_ZExt(m_Value(X)))) && match(Op1, m_One())) {
+      auto *Cttz = IC.Builder.CreateBinaryIntrinsic(Intrinsic::cttz, X,
+                                                     IC.Builder.getTrue());
+      auto *ZextCttz = IC.Builder.CreateZExt(Cttz, II.getType());
+      return IC.replaceInstUsesWith(II, ZextCttz);
+    }
+
     // cttz(abs(x)) -> cttz(x)
     // cttz(nabs(x)) -> cttz(x)
     Value *Y;


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D101582.341680.patch
Type: text/x-patch
Size: 3734 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210429/7d18cf86/attachment.bin>