From 998d22c33264f03e80c8c4ccde5fcafe05441f7b Mon Sep 17 00:00:00 2001
From: Simon Pilgrim <llvm-dev@redking.me.uk>
Date: Sat, 13 Feb 2021 15:01:48 +0000
Subject: [PATCH] [DAG] Fold i1/vXi1 saddsat/uaddsat(x,y) -> or(x,y)

Alive2: https://alive2.llvm.org/ce/z/FzcrpH
---
 lib/CodeGen/SelectionDAG/DAGCombiner.cpp |  4 +
 test/CodeGen/AArch64/sadd_sat_vec.ll     |  9 +--
 test/CodeGen/AArch64/uadd_sat_vec.ll     |  8 +-
 test/CodeGen/X86/avx512-mask-op.ll       | 94 ++++--------------------
 test/CodeGen/X86/sadd_sat_vec.ll         | 60 ++-------------
 test/CodeGen/X86/uadd_sat_vec.ll         | 59 ++-------------
 6 files changed, 32 insertions(+), 202 deletions(-)

diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 42988688111..31d78a6fd67 100644
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -2540,6 +2540,10 @@ SDValue DAGCombiner::visitADDSAT(SDNode *N) {
   if (isNullConstant(N1))
     return N0;
 
+  // fold (add_sat x, y) -> (or x, y) for bool types.
+  if (VT.getScalarType() == MVT::i1)
+    return DAG.getNode(ISD::OR, DL, VT, N0, N1);
+
   // If it cannot overflow, transform into an add.
   if (Opcode == ISD::UADDSAT)
     if (DAG.computeOverflowKind(N0, N1) == SelectionDAG::OFK_Never)
diff --git a/test/CodeGen/AArch64/sadd_sat_vec.ll b/test/CodeGen/AArch64/sadd_sat_vec.ll
index 1ae1ee43bee..2f3cae1aa07 100644
--- a/test/CodeGen/AArch64/sadd_sat_vec.ll
+++ b/test/CodeGen/AArch64/sadd_sat_vec.ll
@@ -286,14 +286,7 @@ define <16 x i4> @v16i4(<16 x i4> %x, <16 x i4> %y) nounwind {
 define <16 x i1> @v16i1(<16 x i1> %x, <16 x i1> %y) nounwind {
 ; CHECK-LABEL: v16i1:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    shl v0.16b, v0.16b, #7
-; CHECK-NEXT:    shl v1.16b, v1.16b, #7
-; CHECK-NEXT:    sshr v0.16b, v0.16b, #7
-; CHECK-NEXT:    sshr v1.16b, v1.16b, #7
-; CHECK-NEXT:    shl v1.16b, v1.16b, #7
-; CHECK-NEXT:    shl v0.16b, v0.16b, #7
-; CHECK-NEXT:    sqadd v0.16b, v0.16b, v1.16b
-; CHECK-NEXT:    sshr v0.16b, v0.16b, #7
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
 ; CHECK-NEXT:    ret
   %z = call <16 x i1> @llvm.sadd.sat.v16i1(<16 x i1> %x, <16 x i1> %y)
   ret <16 x i1> %z
diff --git a/test/CodeGen/AArch64/uadd_sat_vec.ll b/test/CodeGen/AArch64/uadd_sat_vec.ll
index 5f92f713573..43a32b43b58 100644
--- a/test/CodeGen/AArch64/uadd_sat_vec.ll
+++ b/test/CodeGen/AArch64/uadd_sat_vec.ll
@@ -285,13 +285,7 @@ define <16 x i4> @v16i4(<16 x i4> %x, <16 x i4> %y) nounwind {
 define <16 x i1> @v16i1(<16 x i1> %x, <16 x i1> %y) nounwind {
 ; CHECK-LABEL: v16i1:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    movi v2.16b, #1
-; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
-; CHECK-NEXT:    and v1.16b, v1.16b, v2.16b
-; CHECK-NEXT:    shl v1.16b, v1.16b, #7
-; CHECK-NEXT:    shl v0.16b, v0.16b, #7
-; CHECK-NEXT:    uqadd v0.16b, v0.16b, v1.16b
-; CHECK-NEXT:    ushr v0.16b, v0.16b, #7
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
 ; CHECK-NEXT:    ret
   %z = call <16 x i1> @llvm.uadd.sat.v16i1(<16 x i1> %x, <16 x i1> %y)
   ret <16 x i1> %z
diff --git a/test/CodeGen/X86/avx512-mask-op.ll b/test/CodeGen/X86/avx512-mask-op.ll
index 684bebaa85d..9b81809962a 100644
--- a/test/CodeGen/X86/avx512-mask-op.ll
+++ b/test/CodeGen/X86/avx512-mask-op.ll
@@ -5472,48 +5472,17 @@ define i1 @test_v1i1_mul(i1 %x, i1 %y) {
 }
 
 define <1 x i1> @uadd_sat_v1i1(<1 x i1> %x, <1 x i1> %y) nounwind {
-; KNL-LABEL: uadd_sat_v1i1:
-; KNL:       ## %bb.0:
-; KNL-NEXT:    kmovw %edi, %k0
-; KNL-NEXT:    kmovw %esi, %k1
-; KNL-NEXT:    korw %k1, %k0, %k0
-; KNL-NEXT:    kmovw %k0, %eax
-; KNL-NEXT:    ## kill: def $al killed $al killed $eax
-; KNL-NEXT:    retq
-;
-; SKX-LABEL: uadd_sat_v1i1:
-; SKX:       ## %bb.0:
-; SKX-NEXT:    kmovd %edi, %k0
-; SKX-NEXT:    kmovd %esi, %k1
-; SKX-NEXT:    korw %k1, %k0, %k0
-; SKX-NEXT:    kmovd %k0, %eax
-; SKX-NEXT:    ## kill: def $al killed $al killed $eax
-; SKX-NEXT:    retq
-;
-; AVX512BW-LABEL: uadd_sat_v1i1:
-; AVX512BW:       ## %bb.0:
-; AVX512BW-NEXT:    kmovd %edi, %k0
-; AVX512BW-NEXT:    kmovd %esi, %k1
-; AVX512BW-NEXT:    korw %k1, %k0, %k0
-; AVX512BW-NEXT:    kmovd %k0, %eax
-; AVX512BW-NEXT:    ## kill: def $al killed $al killed $eax
-; AVX512BW-NEXT:    retq
-;
-; AVX512DQ-LABEL: uadd_sat_v1i1:
-; AVX512DQ:       ## %bb.0:
-; AVX512DQ-NEXT:    kmovw %edi, %k0
-; AVX512DQ-NEXT:    kmovw %esi, %k1
-; AVX512DQ-NEXT:    korw %k1, %k0, %k0
-; AVX512DQ-NEXT:    kmovw %k0, %eax
-; AVX512DQ-NEXT:    ## kill: def $al killed $al killed $eax
-; AVX512DQ-NEXT:    retq
+; CHECK-LABEL: uadd_sat_v1i1:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    movl %edi, %eax
+; CHECK-NEXT:    orl %esi, %eax
+; CHECK-NEXT:    ## kill: def $al killed $al killed $eax
+; CHECK-NEXT:    retq
 ;
 ; X86-LABEL: uadd_sat_v1i1:
 ; X86:       ## %bb.0:
-; X86-NEXT:    kmovd {{[0-9]+}}(%esp), %k0
-; X86-NEXT:    kmovd {{[0-9]+}}(%esp), %k1
-; X86-NEXT:    korw %k1, %k0, %k0
-; X86-NEXT:    kmovd %k0, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    orl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    ## kill: def $al killed $al killed $eax
 ; X86-NEXT:    retl
   %z = call <1 x i1> @llvm.uadd.sat.v1i1(<1 x i1> %x, <1 x i1> %y)
@@ -5572,48 +5541,17 @@ define <1 x i1> @usub_sat_v1i1(<1 x i1> %x, <1 x i1> %y) nounwind {
 declare <1 x i1> @llvm.usub.sat.v1i1(<1 x i1> %x, <1 x i1> %y)
 
 define <1 x i1> @sadd_sat_v1i1(<1 x i1> %x, <1 x i1> %y) nounwind {
-; KNL-LABEL: sadd_sat_v1i1:
-; KNL:       ## %bb.0:
-; KNL-NEXT:    kmovw %edi, %k0
-; KNL-NEXT:    kmovw %esi, %k1
-; KNL-NEXT:    korw %k1, %k0, %k0
-; KNL-NEXT:    kmovw %k0, %eax
-; KNL-NEXT:    ## kill: def $al killed $al killed $eax
-; KNL-NEXT:    retq
-;
-; SKX-LABEL: sadd_sat_v1i1:
-; SKX:       ## %bb.0:
-; SKX-NEXT:    kmovd %edi, %k0
-; SKX-NEXT:    kmovd %esi, %k1
-; SKX-NEXT:    korw %k1, %k0, %k0
-; SKX-NEXT:    kmovd %k0, %eax
-; SKX-NEXT:    ## kill: def $al killed $al killed $eax
-; SKX-NEXT:    retq
-;
-; AVX512BW-LABEL: sadd_sat_v1i1:
-; AVX512BW:       ## %bb.0:
-; AVX512BW-NEXT:    kmovd %edi, %k0
-; AVX512BW-NEXT:    kmovd %esi, %k1
-; AVX512BW-NEXT:    korw %k1, %k0, %k0
-; AVX512BW-NEXT:    kmovd %k0, %eax
-; AVX512BW-NEXT:    ## kill: def $al killed $al killed $eax
-; AVX512BW-NEXT:    retq
-;
-; AVX512DQ-LABEL: sadd_sat_v1i1:
-; AVX512DQ:       ## %bb.0:
-; AVX512DQ-NEXT:    kmovw %edi, %k0
-; AVX512DQ-NEXT:    kmovw %esi, %k1
-; AVX512DQ-NEXT:    korw %k1, %k0, %k0
-; AVX512DQ-NEXT:    kmovw %k0, %eax
-; AVX512DQ-NEXT:    ## kill: def $al killed $al killed $eax
-; AVX512DQ-NEXT:    retq
+; CHECK-LABEL: sadd_sat_v1i1:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    movl %edi, %eax
+; CHECK-NEXT:    orl %esi, %eax
+; CHECK-NEXT:    ## kill: def $al killed $al killed $eax
+; CHECK-NEXT:    retq
 ;
 ; X86-LABEL: sadd_sat_v1i1:
 ; X86:       ## %bb.0:
-; X86-NEXT:    kmovd {{[0-9]+}}(%esp), %k0
-; X86-NEXT:    kmovd {{[0-9]+}}(%esp), %k1
-; X86-NEXT:    korw %k1, %k0, %k0
-; X86-NEXT:    kmovd %k0, %eax
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    orl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    ## kill: def $al killed $al killed $eax
 ; X86-NEXT:    retl
   %z = call <1 x i1> @llvm.sadd.sat.v1i1(<1 x i1> %x, <1 x i1> %y)
diff --git a/test/CodeGen/X86/sadd_sat_vec.ll b/test/CodeGen/X86/sadd_sat_vec.ll
index d5fdfe9b4ae..9eacc459b3f 100644
--- a/test/CodeGen/X86/sadd_sat_vec.ll
+++ b/test/CodeGen/X86/sadd_sat_vec.ll
@@ -579,63 +579,13 @@ define <16 x i4> @v16i4(<16 x i4> %x, <16 x i4> %y) nounwind {
 define <16 x i1> @v16i1(<16 x i1> %x, <16 x i1> %y) nounwind {
 ; SSE-LABEL: v16i1:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    psllw $7, %xmm1
-; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; SSE-NEXT:    pand %xmm2, %xmm1
-; SSE-NEXT:    psllw $7, %xmm0
-; SSE-NEXT:    pand %xmm2, %xmm0
-; SSE-NEXT:    paddsb %xmm1, %xmm0
-; SSE-NEXT:    pxor %xmm1, %xmm1
-; SSE-NEXT:    pcmpgtb %xmm0, %xmm1
-; SSE-NEXT:    movdqa %xmm1, %xmm0
+; SSE-NEXT:    orps %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: v16i1:
-; AVX1:       # %bb.0:
-; AVX1-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX1-NEXT:    vmovdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; AVX1-NEXT:    vpand %xmm2, %xmm1, %xmm1
-; AVX1-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX1-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpaddsb %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpxor %xmm1, %xmm1, %xmm1
-; AVX1-NEXT:    vpcmpgtb %xmm0, %xmm1, %xmm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: v16i1:
-; AVX2:       # %bb.0:
-; AVX2-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX2-NEXT:    vmovdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; AVX2-NEXT:    vpand %xmm2, %xmm1, %xmm1
-; AVX2-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX2-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; AVX2-NEXT:    vpaddsb %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpxor %xmm1, %xmm1, %xmm1
-; AVX2-NEXT:    vpcmpgtb %xmm0, %xmm1, %xmm0
-; AVX2-NEXT:    retq
-;
-; AVX512F-LABEL: v16i1:
-; AVX512F:       # %bb.0:
-; AVX512F-NEXT:    vpmovsxbd %xmm1, %zmm1
-; AVX512F-NEXT:    vpslld $31, %zmm1, %zmm1
-; AVX512F-NEXT:    vpmovsxbd %xmm0, %zmm0
-; AVX512F-NEXT:    vpslld $31, %zmm0, %zmm0
-; AVX512F-NEXT:    vpord %zmm1, %zmm0, %zmm0
-; AVX512F-NEXT:    vptestmd %zmm0, %zmm0, %k1
-; AVX512F-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0 {%k1} {z}
-; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
-; AVX512F-NEXT:    vzeroupper
-; AVX512F-NEXT:    retq
-;
-; AVX512BW-LABEL: v16i1:
-; AVX512BW:       # %bb.0:
-; AVX512BW-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX512BW-NEXT:    vpmovb2m %xmm1, %k0
-; AVX512BW-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX512BW-NEXT:    vpmovb2m %xmm0, %k1
-; AVX512BW-NEXT:    korw %k0, %k1, %k0
-; AVX512BW-NEXT:    vpmovm2b %k0, %xmm0
-; AVX512BW-NEXT:    retq
+; AVX-LABEL: v16i1:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vorps %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    retq
   %z = call <16 x i1> @llvm.sadd.sat.v16i1(<16 x i1> %x, <16 x i1> %y)
   ret <16 x i1> %z
 }
diff --git a/test/CodeGen/X86/uadd_sat_vec.ll b/test/CodeGen/X86/uadd_sat_vec.ll
index 633238f0b1e..5624dfd8084 100644
--- a/test/CodeGen/X86/uadd_sat_vec.ll
+++ b/test/CodeGen/X86/uadd_sat_vec.ll
@@ -509,62 +509,13 @@ define <16 x i4> @v16i4(<16 x i4> %x, <16 x i4> %y) nounwind {
 define <16 x i1> @v16i1(<16 x i1> %x, <16 x i1> %y) nounwind {
 ; SSE-LABEL: v16i1:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    psllw $7, %xmm1
-; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; SSE-NEXT:    pand %xmm2, %xmm1
-; SSE-NEXT:    psllw $7, %xmm0
-; SSE-NEXT:    pand %xmm2, %xmm0
-; SSE-NEXT:    paddusb %xmm1, %xmm0
-; SSE-NEXT:    psrlw $7, %xmm0
-; SSE-NEXT:    pand {{.*}}(%rip), %xmm0
+; SSE-NEXT:    orps %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
-; AVX1-LABEL: v16i1:
-; AVX1:       # %bb.0:
-; AVX1-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX1-NEXT:    vmovdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; AVX1-NEXT:    vpand %xmm2, %xmm1, %xmm1
-; AVX1-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX1-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpaddusb %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpsrlw $7, %xmm0, %xmm0
-; AVX1-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: v16i1:
-; AVX2:       # %bb.0:
-; AVX2-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX2-NEXT:    vmovdqa {{.*#+}} xmm2 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128]
-; AVX2-NEXT:    vpand %xmm2, %xmm1, %xmm1
-; AVX2-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX2-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; AVX2-NEXT:    vpaddusb %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpsrlw $7, %xmm0, %xmm0
-; AVX2-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
-; AVX2-NEXT:    retq
-;
-; AVX512F-LABEL: v16i1:
-; AVX512F:       # %bb.0:
-; AVX512F-NEXT:    vpmovsxbd %xmm1, %zmm1
-; AVX512F-NEXT:    vpslld $31, %zmm1, %zmm1
-; AVX512F-NEXT:    vpmovsxbd %xmm0, %zmm0
-; AVX512F-NEXT:    vpslld $31, %zmm0, %zmm0
-; AVX512F-NEXT:    vpord %zmm1, %zmm0, %zmm0
-; AVX512F-NEXT:    vptestmd %zmm0, %zmm0, %k1
-; AVX512F-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0 {%k1} {z}
-; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
-; AVX512F-NEXT:    vzeroupper
-; AVX512F-NEXT:    retq
-;
-; AVX512BW-LABEL: v16i1:
-; AVX512BW:       # %bb.0:
-; AVX512BW-NEXT:    vpsllw $7, %xmm1, %xmm1
-; AVX512BW-NEXT:    vpmovb2m %xmm1, %k0
-; AVX512BW-NEXT:    vpsllw $7, %xmm0, %xmm0
-; AVX512BW-NEXT:    vpmovb2m %xmm0, %k1
-; AVX512BW-NEXT:    korw %k0, %k1, %k0
-; AVX512BW-NEXT:    vpmovm2b %k0, %xmm0
-; AVX512BW-NEXT:    retq
+; AVX-LABEL: v16i1:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vorps %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    retq
   %z = call <16 x i1> @llvm.uadd.sat.v16i1(<16 x i1> %x, <16 x i1> %y)
   ret <16 x i1> %z
 }