[llvm] eee993a - [AArch64][SVE] Fold predicate into compare

[Cullen Rhodes via llvm-commits]

Author: Cullen Rhodes
Date: 2022-01-10T10:52:06Z
New Revision: eee993ae4cc77a00681141b044b20ebd06d2aad2

URL: https://github.com/llvm/llvm-project/commit/eee993ae4cc77a00681141b044b20ebd06d2aad2
DIFF: https://github.com/llvm/llvm-project/commit/eee993ae4cc77a00681141b044b20ebd06d2aad2.diff

LOG: [AArch64][SVE] Fold predicate into compare

Codegen of added testcase before this patch:

  ptrue   p0.s
  cmpgt   p1.s, p0/z, z0.s, z1.s
  cmpge   p2.s, p0/z, z2.s, z1.s
  and     p0.b, p0/z, p1.b, p2.b
  ret

Patterns originally authored by Will Lovett.

Reviewed By: david-arm

Differential Revision: https://reviews.llvm.org/D116749

Added: 
    

Modified: 
    llvm/lib/Target/AArch64/SVEInstrFormats.td
    llvm/test/CodeGen/AArch64/sve-intrinsics-int-compares.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AArch64/SVEInstrFormats.td b/llvm/lib/Target/AArch64/SVEInstrFormats.td
index bb488cd7da329..466c785557805 100644
--- a/llvm/lib/Target/AArch64/SVEInstrFormats.td
+++ b/llvm/lib/Target/AArch64/SVEInstrFormats.td
@@ -4641,6 +4641,10 @@ multiclass SVE_SETCC_Pat<CondCode cc, CondCode invcc, ValueType predvt,
             (cmp $Op1, $Op2, $Op3)>;
   def : Pat<(predvt (AArch64setcc_z predvt:$Op1, intvt:$Op2, intvt:$Op3, invcc)),
             (cmp $Op1, $Op3, $Op2)>;
+  def : Pat<(predvt (and predvt:$Pg, (AArch64setcc_z (predvt (AArch64ptrue 31)), intvt:$Op2, intvt:$Op3, cc))),
+            (cmp $Pg, $Op2, $Op3)>;
+  def : Pat<(predvt (and predvt:$Pg, (AArch64setcc_z (predvt (AArch64ptrue 31)), intvt:$Op2, intvt:$Op3, invcc))),
+            (cmp $Pg, $Op3, $Op2)>;
 }
 
 multiclass SVE_SETCC_Pat_With_Zero<CondCode cc, CondCode invcc, ValueType predvt,

diff  --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-int-compares.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-int-compares.ll
index eb885035307ee..9dc807462d913 100644
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-int-compares.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-int-compares.ll
@@ -963,6 +963,39 @@ define <vscale x 4 x i1> @cmpls_wide_splat_s(<vscale x 4 x i1> %pg, <vscale x 4
   ret <vscale x 4 x i1> %out
 }
 
+; Verify general predicate is folded into the compare
+define <vscale x 4 x i1> @predicated_icmp(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) {
+; CHECK-LABEL: predicated_icmp:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    cmpgt p0.s, p0/z, z0.s, z1.s
+; CHECK-NEXT:    cmpge p0.s, p0/z, z2.s, z1.s
+; CHECK-NEXT:    ret
+  %icmp1 = icmp sgt <vscale x 4 x i32> %a, %b
+  %icmp2 = icmp sle <vscale x 4 x i32> %b, %c
+  %and = and <vscale x 4 x i1> %icmp1, %icmp2
+  ret <vscale x 4 x i1> %and
+}
+
+define <vscale x 4 x i1> @predicated_icmp_unknown_lhs(<vscale x 4 x i1> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) {
+; CHECK-LABEL: predicated_icmp_unknown_lhs:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    cmpge p0.s, p0/z, z1.s, z0.s
+; CHECK-NEXT:    ret
+  %icmp = icmp sle <vscale x 4 x i32> %b, %c
+  %and = and <vscale x 4 x i1> %a, %icmp
+  ret <vscale x 4 x i1> %and
+}
+
+define <vscale x 4 x i1> @predicated_icmp_unknown_rhs(<vscale x 4 x i1> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) {
+; CHECK-LABEL: predicated_icmp_unknown_rhs:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    cmpge p0.s, p0/z, z1.s, z0.s
+; CHECK-NEXT:    ret
+  %icmp = icmp sle <vscale x 4 x i32> %b, %c
+  %and = and <vscale x 4 x i1> %icmp, %a
+  ret <vscale x 4 x i1> %and
+}
 
 declare <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
 declare <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)