summaryrefslogtreecommitdiffstats
path: root/test/Transforms/BBVectorize
diff options
context:
space:
mode:
Diffstat (limited to 'test/Transforms/BBVectorize')
-rw-r--r--test/Transforms/BBVectorize/metadata.ll49
-rw-r--r--test/Transforms/BBVectorize/simple-sel.ll29
-rw-r--r--test/Transforms/BBVectorize/simple-tst.ll18
-rw-r--r--test/Transforms/BBVectorize/simple.ll3
-rw-r--r--test/Transforms/BBVectorize/simple3.ll35
5 files changed, 132 insertions, 2 deletions
diff --git a/test/Transforms/BBVectorize/metadata.ll b/test/Transforms/BBVectorize/metadata.ll
new file mode 100644
index 0000000..1e3aaa1
--- /dev/null
+++ b/test/Transforms/BBVectorize/metadata.ll
@@ -0,0 +1,49 @@
+target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
+; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -S | FileCheck %s
+
+; Simple 3-pair chain with loads and stores (with fpmath)
+define void @test1(double* %a, double* %b, double* %c) nounwind uwtable readonly {
+entry:
+ %i0 = load double* %a, align 8
+ %i1 = load double* %b, align 8
+ %mul = fmul double %i0, %i1, !fpmath !2
+ %arrayidx3 = getelementptr inbounds double* %a, i64 1
+ %i3 = load double* %arrayidx3, align 8
+ %arrayidx4 = getelementptr inbounds double* %b, i64 1
+ %i4 = load double* %arrayidx4, align 8
+ %mul5 = fmul double %i3, %i4, !fpmath !3
+ store double %mul, double* %c, align 8
+ %arrayidx5 = getelementptr inbounds double* %c, i64 1
+ store double %mul5, double* %arrayidx5, align 8
+ ret void
+; CHECK: @test1
+; CHECK: !fpmath
+; CHECK: ret void
+}
+
+; Simple 3-pair chain with loads and stores (ints with range)
+define void @test2(i64* %a, i64* %b, i64* %c) nounwind uwtable readonly {
+entry:
+ %i0 = load i64* %a, align 8, !range !0
+ %i1 = load i64* %b, align 8
+ %mul = mul i64 %i0, %i1
+ %arrayidx3 = getelementptr inbounds i64* %a, i64 1
+ %i3 = load i64* %arrayidx3, align 8, !range !1
+ %arrayidx4 = getelementptr inbounds i64* %b, i64 1
+ %i4 = load i64* %arrayidx4, align 8
+ %mul5 = mul i64 %i3, %i4
+ store i64 %mul, i64* %c, align 8
+ %arrayidx5 = getelementptr inbounds i64* %c, i64 1
+ store i64 %mul5, i64* %arrayidx5, align 8
+ ret void
+; CHECK: @test2
+; CHECK-NOT: !range
+; CHECK: ret void
+}
+
+!0 = metadata !{i64 0, i64 2}
+!1 = metadata !{i64 3, i64 5}
+
+!2 = metadata !{ float 5.0 }
+!3 = metadata !{ float 2.5 }
+
diff --git a/test/Transforms/BBVectorize/simple-sel.ll b/test/Transforms/BBVectorize/simple-sel.ll
index 4daa571..325792a 100644
--- a/test/Transforms/BBVectorize/simple-sel.ll
+++ b/test/Transforms/BBVectorize/simple-sel.ll
@@ -1,5 +1,6 @@
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s
+; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -bb-vectorize-no-bools -instcombine -gvn -S | FileCheck %s -check-prefix=CHECK-NB
; Basic depth-3 chain with select
define double @test1(double %A1, double %A2, double %B1, double %B2, i1 %C1, i1 %C2) {
@@ -27,4 +28,32 @@ define double @test1(double %A1, double %A2, double %B1, double %B2, i1 %C1, i1
; CHECK: ret double %R
}
+; Basic depth-3 chain with select (and vect. compare)
+define double @test2(double %A1, double %A2, double %B1, double %B2) {
+; CHECK: @test2
+; CHECK-NB: @test2
+; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0
+; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0
+; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1
+; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1
+ %X1 = fsub double %A1, %B1
+ %X2 = fsub double %A2, %B2
+; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2
+ %Y1 = fmul double %X1, %A1
+ %Y2 = fmul double %X2, %A2
+; CHECK: %Y1 = fmul <2 x double> %X1, %X1.v.i0.2
+ %C1 = fcmp ogt double %X1, %A1
+ %C2 = fcmp ogt double %X2, %A2
+; CHECK: %C1 = fcmp ogt <2 x double> %X1, %X1.v.i0.2
+; CHECK-NB: fcmp ogt double
+ %Z1 = select i1 %C1, double %Y1, double %B1
+ %Z2 = select i1 %C2, double %Y2, double %B2
+; CHECK: %Z1 = select <2 x i1> %C1, <2 x double> %Y1, <2 x double> %X1.v.i1.2
+ %R = fmul double %Z1, %Z2
+; CHECK: %Z1.v.r1 = extractelement <2 x double> %Z1, i32 0
+; CHECK: %Z1.v.r2 = extractelement <2 x double> %Z1, i32 1
+; CHECK: %R = fmul double %Z1.v.r1, %Z1.v.r2
+ ret double %R
+; CHECK: ret double %R
+}
diff --git a/test/Transforms/BBVectorize/simple-tst.ll b/test/Transforms/BBVectorize/simple-tst.ll
new file mode 100644
index 0000000..42146c6
--- /dev/null
+++ b/test/Transforms/BBVectorize/simple-tst.ll
@@ -0,0 +1,18 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux"
+; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -bb-vectorize-vector-bits=256 -instcombine -gvn -S | FileCheck %s
+
+; Basic depth-3 chain (target-specific type should not vectorize)
+define ppc_fp128 @test7(ppc_fp128 %A1, ppc_fp128 %A2, ppc_fp128 %B1, ppc_fp128 %B2) {
+; CHECK: @test7
+; CHECK-NOT: <2 x ppc_fp128>
+ %X1 = fsub ppc_fp128 %A1, %B1
+ %X2 = fsub ppc_fp128 %A2, %B2
+ %Y1 = fmul ppc_fp128 %X1, %A1
+ %Y2 = fmul ppc_fp128 %X2, %A2
+ %Z1 = fadd ppc_fp128 %Y1, %B1
+ %Z2 = fadd ppc_fp128 %Y2, %B2
+ %R = fmul ppc_fp128 %Z1, %Z2
+ ret ppc_fp128 %R
+}
+
diff --git a/test/Transforms/BBVectorize/simple.ll b/test/Transforms/BBVectorize/simple.ll
index 904d766..88eb9c9 100644
--- a/test/Transforms/BBVectorize/simple.ll
+++ b/test/Transforms/BBVectorize/simple.ll
@@ -138,8 +138,7 @@ define <8 x i8> @test6(<8 x i8> %A1, <8 x i8> %A2, <8 x i8> %B1, <8 x i8> %B2) {
; CHECK: %Z1 = add <16 x i8> %Y1, %X1.v.i1
%Q1 = shufflevector <8 x i8> %Z1, <8 x i8> %Z2, <8 x i32> <i32 15, i32 8, i32 6, i32 1, i32 13, i32 10, i32 4, i32 3>
%Q2 = shufflevector <8 x i8> %Z2, <8 x i8> %Z2, <8 x i32> <i32 6, i32 7, i32 0, i32 1, i32 2, i32 4, i32 4, i32 1>
-; CHECK: %Z1.v.r2 = shufflevector <16 x i8> %Z1, <16 x i8> undef, <8 x i32> <i32 8, i32 undef, i32 10, i32 undef, i32 undef, i32 13, i32 undef, i32 15>
-; CHECK: %Q1.v.i1 = shufflevector <8 x i8> %Z1.v.r2, <8 x i8> undef, <16 x i32> <i32 0, i32 undef, i32 2, i32 undef, i32 undef, i32 5, i32 undef, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK: %Q1.v.i1 = shufflevector <16 x i8> %Z1, <16 x i8> undef, <16 x i32> <i32 8, i32 undef, i32 10, i32 undef, i32 undef, i32 13, i32 undef, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
; CHECK: %Q1 = shufflevector <16 x i8> %Z1, <16 x i8> %Q1.v.i1, <16 x i32> <i32 23, i32 16, i32 6, i32 1, i32 21, i32 18, i32 4, i32 3, i32 14, i32 15, i32 8, i32 9, i32 10, i32 12, i32 12, i32 9>
%R = mul <8 x i8> %Q1, %Q2
; CHECK: %Q1.v.r1 = shufflevector <16 x i8> %Q1, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
diff --git a/test/Transforms/BBVectorize/simple3.ll b/test/Transforms/BBVectorize/simple3.ll
new file mode 100644
index 0000000..153be73
--- /dev/null
+++ b/test/Transforms/BBVectorize/simple3.ll
@@ -0,0 +1,35 @@
+target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
+; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -bb-vectorize-vector-bits=192 -instcombine -gvn -S | FileCheck %s
+
+; Basic depth-3 chain
+define double @test1(double %A1, double %A2, double %A3, double %B1, double %B2, double %B3) {
+; CHECK: @test1
+; CHECK: %X1.v.i1.11 = insertelement <3 x double> undef, double %B1, i32 0
+; CHECK: %X1.v.i1.22 = insertelement <3 x double> %X1.v.i1.11, double %B2, i32 1
+; CHECK: %X1.v.i1 = insertelement <3 x double> %X1.v.i1.22, double %B3, i32 2
+; CHECK: %X1.v.i0.13 = insertelement <3 x double> undef, double %A1, i32 0
+; CHECK: %X1.v.i0.24 = insertelement <3 x double> %X1.v.i0.13, double %A2, i32 1
+; CHECK: %X1.v.i0 = insertelement <3 x double> %X1.v.i0.24, double %A3, i32 2
+ %X1 = fsub double %A1, %B1
+ %X2 = fsub double %A2, %B2
+ %X3 = fsub double %A3, %B3
+; CHECK: %X1 = fsub <3 x double> %X1.v.i0, %X1.v.i1
+ %Y1 = fmul double %X1, %A1
+ %Y2 = fmul double %X2, %A2
+ %Y3 = fmul double %X3, %A3
+; CHECK: %Y1 = fmul <3 x double> %X1, %X1.v.i0
+ %Z1 = fadd double %Y1, %B1
+ %Z2 = fadd double %Y2, %B2
+ %Z3 = fadd double %Y3, %B3
+; CHECK: %Z1 = fadd <3 x double> %Y1, %X1.v.i1
+ %R1 = fmul double %Z1, %Z2
+ %R = fmul double %R1, %Z3
+; CHECK: %Z1.v.r210 = extractelement <3 x double> %Z1, i32 2
+; CHECK: %Z1.v.r1 = extractelement <3 x double> %Z1, i32 0
+; CHECK: %Z1.v.r2 = extractelement <3 x double> %Z1, i32 1
+; CHECK: %R1 = fmul double %Z1.v.r1, %Z1.v.r2
+; CHECK: %R = fmul double %R1, %Z1.v.r210
+ ret double %R
+; CHECK: ret double %R
+}
+
OpenPOWER on IntegriCloud