[Power9] add more tests for D38287; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@319328 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/test/CodeGen/PowerPC/vec_extract_p9_2.ll b/test/CodeGen/PowerPC/vec_extract_p9_2.ll
new file mode 100644 (file)
index 0000000..9734a88
--- /dev/null
+++ b/test/CodeGen/PowerPC/vec_extract_p9_2.ll
@@ -0,0 +1,248 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-gnu-linux -mcpu=pwr9 < %s | FileCheck %s -check-prefix=CHECK-LE
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-gnu-linux -mcpu=pwr9 < %s | FileCheck %s -check-prefix=CHECK-BE
+
+define zeroext i8 @test_add1(<16 x i8> %a, i32 signext %index, i8 zeroext %c) {
+; CHECK-LE-LABEL: test_add1:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    vextubrx 3, 5, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    rlwinm 3, 3, 0, 24, 31
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add1:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    vextublx 3, 5, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    rlwinm 3, 3, 0, 24, 31
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <16 x i8> %a, i32 %index
+  %conv = zext i8 %vecext to i32
+  %conv1 = zext i8 %c to i32
+  %add = add nuw nsw i32 %conv, %conv1
+  %conv2 = trunc i32 %add to i8
+  ret i8 %conv2
+}
+
+define signext i8 @test_add2(<16 x i8> %a, i32 signext %index, i8 signext %c) {
+; CHECK-LE-LABEL: test_add2:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    vextubrx 3, 5, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    extsb 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add2:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    vextublx 3, 5, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    extsb 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <16 x i8> %a, i32 %index
+  %conv3 = zext i8 %vecext to i32
+  %conv14 = zext i8 %c to i32
+  %add = add nuw nsw i32 %conv3, %conv14
+  %conv2 = trunc i32 %add to i8
+  ret i8 %conv2
+}
+
+define zeroext i16 @test_add3(<8 x i16> %a, i32 signext %index, i16 zeroext %c) {
+; CHECK-LE-LABEL: test_add3:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    rlwinm 3, 5, 1, 28, 30
+; CHECK-LE-NEXT:    vextuhrx 3, 3, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    rlwinm 3, 3, 0, 16, 31
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add3:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    rlwinm 3, 5, 1, 28, 30
+; CHECK-BE-NEXT:    vextuhlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    rlwinm 3, 3, 0, 16, 31
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <8 x i16> %a, i32 %index
+  %conv = zext i16 %vecext to i32
+  %conv1 = zext i16 %c to i32
+  %add = add nuw nsw i32 %conv, %conv1
+  %conv2 = trunc i32 %add to i16
+  ret i16 %conv2
+}
+
+define signext i16 @test_add4(<8 x i16> %a, i32 signext %index, i16 signext %c) {
+; CHECK-LE-LABEL: test_add4:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    rlwinm 3, 5, 1, 28, 30
+; CHECK-LE-NEXT:    vextuhrx 3, 3, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    extsh 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add4:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    rlwinm 3, 5, 1, 28, 30
+; CHECK-BE-NEXT:    vextuhlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    extsh 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <8 x i16> %a, i32 %index
+  %conv5 = zext i16 %vecext to i32
+  %conv16 = zext i16 %c to i32
+  %add = add nuw nsw i32 %conv5, %conv16
+  %conv2 = trunc i32 %add to i16
+  ret i16 %conv2
+}
+
+define zeroext i32 @test_add5(<4 x i32> %a, i32 signext %index, i32 zeroext %c) {
+; CHECK-LE-LABEL: test_add5:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    rlwinm 3, 5, 2, 28, 29
+; CHECK-LE-NEXT:    vextuwrx 3, 3, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    clrldi 3, 3, 32
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add5:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    rlwinm 3, 5, 2, 28, 29
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    clrldi 3, 3, 32
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 %index
+  %add = add i32 %vecext, %c
+  ret i32 %add
+}
+
+define signext i32 @test_add6(<4 x i32> %a, i32 signext %index, i32 signext %c) {
+; CHECK-LE-LABEL: test_add6:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    rlwinm 3, 5, 2, 28, 29
+; CHECK-LE-NEXT:    vextuwrx 3, 3, 2
+; CHECK-LE-NEXT:    add 3, 3, 6
+; CHECK-LE-NEXT:    extsw 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test_add6:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    rlwinm 3, 5, 2, 28, 29
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 6
+; CHECK-BE-NEXT:    extsw 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 %index
+  %add = add nsw i32 %vecext, %c
+  ret i32 %add
+}
+
+; When extracting word element 2 on LE, it's better to use mfvsrwz rather than vextuwrx
+define zeroext i32 @test7(<4 x i32> %a) {
+; CHECK-LE-LABEL: test7:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    mfvsrwz 3, 34
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test7:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    li 3, 8
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 2
+  ret i32 %vecext
+}
+
+define zeroext i32 @testadd_7(<4 x i32> %a, i32 zeroext %c) {
+; CHECK-LE-LABEL: testadd_7:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    mfvsrwz 3, 34
+; CHECK-LE-NEXT:    add 3, 3, 5
+; CHECK-LE-NEXT:    clrldi 3, 3, 32
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: testadd_7:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    li 3, 8
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 5
+; CHECK-BE-NEXT:    clrldi 3, 3, 32
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 2
+  %add = add i32 %vecext, %c
+  ret i32 %add
+}
+
+define signext i32 @test8(<4 x i32> %a) {
+; CHECK-LE-LABEL: test8:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    mfvsrwz 3, 34
+; CHECK-LE-NEXT:    extsw 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test8:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    li 3, 8
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    extsw 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 2
+  ret i32 %vecext
+}
+
+define signext i32 @testadd_8(<4 x i32> %a, i32 signext %c) {
+; CHECK-LE-LABEL: testadd_8:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    mfvsrwz 3, 34
+; CHECK-LE-NEXT:    add 3, 3, 5
+; CHECK-LE-NEXT:    extsw 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: testadd_8:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    li 3, 8
+; CHECK-BE-NEXT:    vextuwlx 3, 3, 2
+; CHECK-BE-NEXT:    add 3, 3, 5
+; CHECK-BE-NEXT:    extsw 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 2
+  %add = add nsw i32 %vecext, %c
+  ret i32 %add
+}
+
+; When extracting word element 1 on BE, it's better to use mfvsrwz rather than vextuwlx
+define signext i32 @test9(<4 x i32> %a) {
+; CHECK-LE-LABEL: test9:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    li 3, 4
+; CHECK-LE-NEXT:    vextuwrx 3, 3, 2
+; CHECK-LE-NEXT:    extsw 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: test9:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    mfvsrwz 3, 34
+; CHECK-BE-NEXT:    extsw 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 1
+  ret i32 %vecext
+}
+
+define signext i32 @testadd_9(<4 x i32> %a, i32 signext %c) {
+; CHECK-LE-LABEL: testadd_9:
+; CHECK-LE:       # BB#0: # %entry
+; CHECK-LE-NEXT:    li 3, 4
+; CHECK-LE-NEXT:    vextuwrx 3, 3, 2
+; CHECK-LE-NEXT:    add 3, 3, 5
+; CHECK-LE-NEXT:    extsw 3, 3
+; CHECK-LE-NEXT:    blr
+; CHECK-BE-LABEL: testadd_9:
+; CHECK-BE:       # BB#0: # %entry
+; CHECK-BE-NEXT:    mfvsrwz 3, 34
+; CHECK-BE-NEXT:    add 3, 3, 5
+; CHECK-BE-NEXT:    extsw 3, 3
+; CHECK-BE-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %a, i32 1
+  %add = add nsw i32 %vecext, %c
+  ret i32 %add
+}