[llvm] [InstCombine] fold `ldexp(x, zext(i1 y))` to `fmul x, (select y, 2.0, 1.0)` (PR #94887)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,46 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes=instcombine -S | FileCheck %s
+
+define float @ldexp_zext_float(float %x, i1 %bool) {
+; CHECK-LABEL: @ldexp_zext_float(
+; CHECK-NEXT:    [[TMP1:%.*]] = select i1 [[BOOL:%.*]], float 2.000000e+00, float 1.000000e+00
+; CHECK-NEXT:    [[LDEXP:%.*]] = fmul float [[TMP1]], [[X:%.*]]
+; CHECK-NEXT:    ret float [[LDEXP]]
+;
+  %zext = zext i1 %bool to i32
+  %ldexp = call float @llvm.ldexp.f32.i32(float %x, i32 %zext)
+  ret float %ldexp
+}
+
+define float @ldexp_zext_float_negative(float %x, i8 %y) {
+; CHECK-LABEL: @ldexp_zext_float_negative(
+; CHECK-NEXT:    [[ZEXT:%.*]] = zext i8 [[Y:%.*]] to i32
+; CHECK-NEXT:    [[LDEXP:%.*]] = call float @llvm.ldexp.f32.i32(float [[X:%.*]], i32 [[ZEXT]])
+; CHECK-NEXT:    ret float [[LDEXP]]
+;
+  %zext = zext i8 %y to i32
+  %ldexp = call float @llvm.ldexp.f32.i32(float %x, i32 %zext)
+  ret float %ldexp
+}
+
+define double @ldexp_zext_double(double %x, i1 %bool) {
+; CHECK-LABEL: @ldexp_zext_double(
+; CHECK-NEXT:    [[TMP1:%.*]] = select i1 [[BOOL:%.*]], double 2.000000e+00, double 1.000000e+00
+; CHECK-NEXT:    [[LDEXP:%.*]] = fmul double [[TMP1]], [[X:%.*]]
+; CHECK-NEXT:    ret double [[LDEXP]]
+;
+  %zext = zext i1 %bool to i32
+  %ldexp = call double @llvm.ldexp.f32.i32(double %x, i32 %zext)
+  ret double %ldexp
+}
+
+define <2 x float> @ldexp_zext_float_vector(<2 x float> %x, <2 x i1> %bool) {
+; CHECK-LABEL: @ldexp_zext_float_vector(
+; CHECK-NEXT:    [[TMP1:%.*]] = select <2 x i1> [[BOOL:%.*]], <2 x float> <float 2.000000e+00, float 2.000000e+00>, <2 x float> <float 1.000000e+00, float 1.000000e+00>
+; CHECK-NEXT:    [[LDEXP:%.*]] = fmul <2 x float> [[TMP1]], [[X:%.*]]
+; CHECK-NEXT:    ret <2 x float> [[LDEXP]]
+;
+  %zext = zext <2 x i1> %bool to <2 x i32>
+  %ldexp = call <2 x float> @llvm.ldexp.v2f32.v2i32(<2 x float> %x, <2 x i32> %zext)
+  ret <2 x float> %ldexp
+}
----------------
arsenm wrote:

Test fast math flag preservation on the ldexp 

https://github.com/llvm/llvm-project/pull/94887