[AArch64][FastISel] Fix integer extend optimization.

[oota-llvm.git] / test / CodeGen / AArch64 / neon-fma.ll

diff --git a/test/CodeGen/AArch64/neon-fma.ll b/test/CodeGen/AArch64/neon-fma.ll
index dcf4e2878068acf6949cac03a2ec4e4627a45be3..af70302ca939c4c3c60e6225c901a0474ad584be 100644 (file)
--- a/test/CodeGen/AArch64/neon-fma.ll
+++ b/test/CodeGen/AArch64/neon-fma.ll
@@ -1,21 +1,21 @@
 ; RUN: llc < %s -verify-machineinstrs -mtriple=aarch64-none-linux-gnu -mattr=+neon -fp-contract=fast | FileCheck %s
 
 define <2 x float> @fmla2xfloat(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.2s, {{v[0-31]+}}.2s, {{v[0-31]+}}.2s
+;CHECK: fmla {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
        %tmp1 = fmul <2 x float> %A, %B;
        %tmp2 = fadd <2 x float> %C, %tmp1;
        ret <2 x float> %tmp2
 }
 
 define <4 x float> @fmla4xfloat(<4 x float> %A, <4 x float> %B, <4 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.4s, {{v[0-31]+}}.4s, {{v[0-31]+}}.4s
+;CHECK: fmla {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
        %tmp1 = fmul <4 x float> %A, %B;
        %tmp2 = fadd <4 x float> %C, %tmp1;
        ret <4 x float> %tmp2
 }
 
 define <2 x double> @fmla2xdouble(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
-;CHECK: fmla {{v[0-31]+}}.2d, {{v[0-31]+}}.2d, {{v[0-31]+}}.2d
+;CHECK: fmla {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
        %tmp1 = fmul <2 x double> %A, %B;
        %tmp2 = fadd <2 x double> %C, %tmp1;
        ret <2 x double> %tmp2
@@ -23,21 +23,21 @@ define <2 x double> @fmla2xdouble(<2 x double> %A, <2 x double> %B, <2 x double>
 
 
 define <2 x float> @fmls2xfloat(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
-;CHECK: fmls {{v[0-31]+}}.2s, {{v[0-31]+}}.2s, {{v[0-31]+}}.2s
+;CHECK: fmls {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
        %tmp1 = fmul <2 x float> %A, %B;
        %tmp2 = fsub <2 x float> %C, %tmp1;
        ret <2 x float> %tmp2
 }
 
 define <4 x float> @fmls4xfloat(<4 x float> %A, <4 x float> %B, <4 x float> %C) {
-;CHECK: fmls {{v[0-31]+}}.4s, {{v[0-31]+}}.4s, {{v[0-31]+}}.4s
+;CHECK: fmls {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
        %tmp1 = fmul <4 x float> %A, %B;
        %tmp2 = fsub <4 x float> %C, %tmp1;
        ret <4 x float> %tmp2
 }
 
 define <2 x double> @fmls2xdouble(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
-;CHECK: fmls {{v[0-31]+}}.2d, {{v[0-31]+}}.2d, {{v[0-31]+}}.2d
+;CHECK: fmls {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
        %tmp1 = fmul <2 x double> %A, %B;
        %tmp2 = fsub <2 x double> %C, %tmp1;
        ret <2 x double> %tmp2
@@ -51,39 +51,39 @@ declare <4 x float> @llvm.fma.v4f32(<4 x float>, <4 x float>, <4 x float>)
 declare <2 x double> @llvm.fma.v2f64(<2 x double>, <2 x double>, <2 x double>)
 
 define <2 x float> @fmla2xfloat_fused(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.2s, {{v[0-31]+}}.2s, {{v[0-31]+}}.2s
+;CHECK: fmla {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
         %val = call <2 x float> @llvm.fma.v2f32(<2 x float> %A, <2 x float> %B, <2 x float> %C)
        ret <2 x float> %val
 }
 
 define <4 x float> @fmla4xfloat_fused(<4 x float> %A, <4 x float> %B, <4 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.4s, {{v[0-31]+}}.4s, {{v[0-31]+}}.4s
+;CHECK: fmla {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
         %val = call <4 x float> @llvm.fma.v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C)
        ret <4 x float> %val
 }
 
 define <2 x double> @fmla2xdouble_fused(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
-;CHECK: fmla {{v[0-31]+}}.2d, {{v[0-31]+}}.2d, {{v[0-31]+}}.2d
+;CHECK: fmla {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
         %val = call <2 x double> @llvm.fma.v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C)
        ret <2 x double> %val
 }
 
 define <2 x float> @fmls2xfloat_fused(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
-;CHECK: fmls {{v[0-31]+}}.2s, {{v[0-31]+}}.2s, {{v[0-31]+}}.2s
+;CHECK: fmls {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
         %negA = fsub <2 x float> <float -0.0, float -0.0>, %A
         %val = call <2 x float> @llvm.fma.v2f32(<2 x float> %negA, <2 x float> %B, <2 x float> %C)
        ret <2 x float> %val
 }
 
 define <4 x float> @fmls4xfloat_fused(<4 x float> %A, <4 x float> %B, <4 x float> %C) {
-;CHECK: fmls {{v[0-31]+}}.4s, {{v[0-31]+}}.4s, {{v[0-31]+}}.4s
+;CHECK: fmls {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
         %negA = fsub <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %A
         %val = call <4 x float> @llvm.fma.v4f32(<4 x float> %negA, <4 x float> %B, <4 x float> %C)
        ret <4 x float> %val
 }
 
 define <2 x double> @fmls2xdouble_fused(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
-;CHECK: fmls {{v[0-31]+}}.2d, {{v[0-31]+}}.2d, {{v[0-31]+}}.2d
+;CHECK: fmls {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
         %negA = fsub <2 x double> <double -0.0, double -0.0>, %A
         %val = call <2 x double> @llvm.fma.v2f64(<2 x double> %negA, <2 x double> %B, <2 x double> %C)
        ret <2 x double> %val
@@ -94,19 +94,39 @@ declare <4 x float> @llvm.fmuladd.v4f32(<4 x float>, <4 x float>, <4 x float>)
 declare <2 x double> @llvm.fmuladd.v2f64(<2 x double>, <2 x double>, <2 x double>)
 
 define <2 x float> @fmuladd2xfloat(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.2s, {{v[0-31]+}}.2s, {{v[0-31]+}}.2s
+;CHECK: fmla {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
         %val = call <2 x float> @llvm.fmuladd.v2f32(<2 x float> %A, <2 x float> %B, <2 x float> %C)
        ret <2 x float> %val
 }
 
 define <4 x float> @fmuladd4xfloat_fused(<4 x float> %A, <4 x float> %B, <4 x float> %C) {
-;CHECK: fmla {{v[0-31]+}}.4s, {{v[0-31]+}}.4s, {{v[0-31]+}}.4s
+;CHECK: fmla {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
         %val = call <4 x float> @llvm.fmuladd.v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C)
        ret <4 x float> %val
 }
 
 define <2 x double> @fmuladd2xdouble_fused(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
-;CHECK: fmla {{v[0-31]+}}.2d, {{v[0-31]+}}.2d, {{v[0-31]+}}.2d
+;CHECK: fmla {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
         %val = call <2 x double> @llvm.fmuladd.v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C)
        ret <2 x double> %val
 }
+
+
+; Another set of tests that check for multiply single use
+
+define <2 x float> @fmla2xfloati_su(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
+;CHECK-NOT: fmla {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
+  %tmp1 = fmul <2 x float> %A, %B;
+  %tmp2 = fadd <2 x float> %C, %tmp1;
+  %tmp3 = fadd <2 x float> %tmp2, %tmp1;
+  ret <2 x float> %tmp3
+}
+
+define <2 x double> @fmls2xdouble_su(<2 x double> %A, <2 x double> %B, <2 x double> %C) {
+;CHECK-NOT: fmls {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
+        %tmp1 = fmul <2 x double> %A, %B;
+        %tmp2 = fsub <2 x double> %C, %tmp1;
+        %tmp3 = fsub <2 x double> %tmp2, %tmp1;
+        ret <2 x double> %tmp3
+}
+