[AArch64][SVE] Detect MOV (imm, pred, zeroing/merging)

llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td

@@ -892,6 +892,26 @@ let Predicates = [HasSVEorSME] in {
   def : Pat<(nxv2i64 (splat_vector (i64 (SVECpyDupImm64Pat i32:$a, i32:$b)))),
             (DUP_ZI_D $a, $b)>;
 
+  // Duplicate Int immediate to active vector elements (zeroing).
+  def : Pat<(nxv16i8 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm8Pat i32:$a, i32:$b)), (SVEDup0Undef))),
+            (CPY_ZPzI_B $pg, $a, $b)>;
+  def : Pat<(nxv8i16 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm16Pat i32:$a, i32:$b)), (SVEDup0Undef))),
+            (CPY_ZPzI_H $pg, $a, $b)>;
+  def : Pat<(nxv4i32 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm32Pat i32:$a, i32:$b)), (SVEDup0Undef))),
+            (CPY_ZPzI_S $pg, $a, $b)>;
+  def : Pat<(nxv2i64 (AArch64dup_mt PPR:$pg, (i64 (SVECpyDupImm64Pat i32:$a, i32:$b)), (SVEDup0Undef))),
+            (CPY_ZPzI_D $pg, $a, $b)>;
+
+  // Duplicate Int immediate to active vector elements (merging).
+  def : Pat<(nxv16i8 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm8Pat i32:$a, i32:$b)), (nxv16i8 ZPR:$z))),
+            (CPY_ZPmI_B $z, $pg, $a, $b)>;
+  def : Pat<(nxv8i16 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm16Pat i32:$a, i32:$b)), (nxv8i16 ZPR:$z))),
+            (CPY_ZPmI_H $z, $pg, $a, $b)>;
+  def : Pat<(nxv4i32 (AArch64dup_mt PPR:$pg, (i32 (SVECpyDupImm32Pat i32:$a, i32:$b)), (nxv4i32 ZPR:$z))),
+            (CPY_ZPmI_S $z, $pg, $a, $b)>;
+  def : Pat<(nxv2i64 (AArch64dup_mt PPR:$pg, (i64 (SVECpyDupImm64Pat i32:$a, i32:$b)), (nxv2i64 ZPR:$z))),
+            (CPY_ZPmI_D $z, $pg, $a, $b)>;
+
   // Duplicate immediate FP into all vector elements.
   def : Pat<(nxv2f16 (splat_vector (f16 fpimm:$val))),
             (DUP_ZR_H (MOVi32imm (bitcast_fpimm_to_i32 f16:$val)))>;

llvm/test/CodeGen/AArch64/sve-mov-imm-pred.ll

@@ -0,0 +1,83 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s
+
+; Zeroing.
+
+define dso_local <vscale x 16 x i8> @mov_z_b(<vscale x 16 x i1> %pg) {
+; CHECK-LABEL: mov_z_b:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.b, p0/z, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, i8 1)
+  ret <vscale x 16 x i8> %r
+}
+
+define dso_local <vscale x 8 x i16> @mov_z_h(<vscale x 8 x i1> %pg) {
+; CHECK-LABEL: mov_z_h:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.h, p0/z, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %pg, i16 1)
+  ret <vscale x 8 x i16> %r
+}
+
+define dso_local <vscale x 4 x i32> @mov_z_s(<vscale x 4 x i1> %pg) {
+; CHECK-LABEL: mov_z_s:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.s, p0/z, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %pg, i32 1)
+  ret <vscale x 4 x i32> %r
+}
+
+define dso_local <vscale x 2 x i64> @mov_z_d(<vscale x 2 x i1> %pg) {
+; CHECK-LABEL: mov_z_d:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.d, p0/z, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %pg, i64 1)
+  ret <vscale x 2 x i64> %r
+}
+
+; Merging.
+
+define dso_local <vscale x 16 x i8> @mov_m_b(<vscale x 16 x i8> %zd, <vscale x 16 x i1> %pg) {
+; CHECK-LABEL: mov_m_b:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.b, p0/m, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> %zd, <vscale x 16 x i1> %pg, i8 1)
+  ret <vscale x 16 x i8> %r
+}
+
+define dso_local <vscale x 8 x i16> @mov_m_h(<vscale x 8 x i16> %zd, <vscale x 8 x i1> %pg) {
+; CHECK-LABEL: mov_m_h:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.h, p0/m, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> %zd, <vscale x 8 x i1> %pg, i16 1)
+  ret <vscale x 8 x i16> %r
+}
+
+define dso_local <vscale x 4 x i32> @mov_m_s(<vscale x 4 x i32> %zd, <vscale x 4 x i1> %pg) {
+; CHECK-LABEL: mov_m_s:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.s, p0/m, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> %zd, <vscale x 4 x i1> %pg, i32 1)
+  ret <vscale x 4 x i32> %r
+}
+
+define dso_local <vscale x 2 x i64> @mov_m_d(<vscale x 2 x i64> %zd, <vscale x 2 x i1> %pg) {
+; CHECK-LABEL: mov_m_d:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    mov z0.d, p0/m, #1 // =0x1
+; CHECK-NEXT:    ret
+  %r = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> %zd, <vscale x 2 x i1> %pg, i64 1)
+  ret <vscale x 2 x i64> %r
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i1>, i8)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i1>, i16)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i1>, i32)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i1>, i64)