[AArch64][SVE] Add SVE intrinsics for address calculations

[kmclaughlin]

Summary

Adds the @llvm.aarch64.sve.adr[b|h|w|d] intrinsics

[sdesmalen]

LGTM!

llvm/include/llvm/IR/IntrinsicsAArch64.td

 // Scalar to vector operations
 //
 
 def int_aarch64_sve_dup : AdvSIMD_SVE_DUP_Intrinsic;
 
 def int_aarch64_sve_index : AdvSIMD_SVE_Index_Intrinsic;
 
 //
+// Address calculation
+//
+
+def int_aarch64_sve_adrb : AdvSIMD_2VectorArg_Intrinsic;
+def int_aarch64_sve_adrh : AdvSIMD_2VectorArg_Intrinsic;
+def int_aarch64_sve_adrw : AdvSIMD_2VectorArg_Intrinsic;
+def int_aarch64_sve_adrd : AdvSIMD_2VectorArg_Intrinsic;
+
+//
 // Integer arithmetic
 //
 
 def int_aarch64_sve_add   : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_sub   : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_subr  : AdvSIMD_Pred2VectorArg_Intrinsic;
 
 def int_aarch64_sve_pmul       : AdvSIMD_2VectorArg_Intrinsic;

llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td

multiclass sve_prefetch<SDPatternOperator prefetch, ValueType PredTy, Instruction RegImmInst, Instruction RegRegInst, int scale, ComplexPattern AddrCP> {
   defm PRFW_D_PZI : sve_mem_64b_prfm_vi<0b10, "prfw", uimm5s4>;
   defm PRFD_D_PZI : sve_mem_64b_prfm_vi<0b11, "prfd", uimm5s8>;
 
   defm ADR_SXTW_ZZZ_D : sve_int_bin_cons_misc_0_a_sxtw<0b00, "adr">;
   defm ADR_UXTW_ZZZ_D : sve_int_bin_cons_misc_0_a_uxtw<0b01, "adr">;
   defm ADR_LSL_ZZZ_S  : sve_int_bin_cons_misc_0_a_32_lsl<0b10, "adr">;
   defm ADR_LSL_ZZZ_D  : sve_int_bin_cons_misc_0_a_64_lsl<0b11, "adr">;
 
+  def : Pat<(nxv4i32 (int_aarch64_sve_adrb nxv4i32:$Op1, nxv4i32:$Op2)),
+            (ADR_LSL_ZZZ_S_0 $Op1, $Op2)>;
+  def : Pat<(nxv4i32 (int_aarch64_sve_adrh nxv4i32:$Op1, nxv4i32:$Op2)),
+            (ADR_LSL_ZZZ_S_1 $Op1, $Op2)>;
+  def : Pat<(nxv4i32 (int_aarch64_sve_adrw nxv4i32:$Op1, nxv4i32:$Op2)),
+            (ADR_LSL_ZZZ_S_2 $Op1, $Op2)>;
+  def : Pat<(nxv4i32 (int_aarch64_sve_adrd nxv4i32:$Op1, nxv4i32:$Op2)),
+            (ADR_LSL_ZZZ_S_3 $Op1, $Op2)>;
+
+  def : Pat<(nxv2i64 (int_aarch64_sve_adrb nxv2i64:$Op1, nxv2i64:$Op2)),
+            (ADR_LSL_ZZZ_D_0 $Op1, $Op2)>;
+  def : Pat<(nxv2i64 (int_aarch64_sve_adrh nxv2i64:$Op1, nxv2i64:$Op2)),
+            (ADR_LSL_ZZZ_D_1 $Op1, $Op2)>;
+  def : Pat<(nxv2i64 (int_aarch64_sve_adrw nxv2i64:$Op1, nxv2i64:$Op2)),
+            (ADR_LSL_ZZZ_D_2 $Op1, $Op2)>;
+  def : Pat<(nxv2i64 (int_aarch64_sve_adrd nxv2i64:$Op1, nxv2i64:$Op2)),
+            (ADR_LSL_ZZZ_D_3 $Op1, $Op2)>;
+
   defm TBL_ZZZ  : sve_int_perm_tbl<"tbl", AArch64tbl>;
 
   defm ZIP1_ZZZ : sve_int_perm_bin_perm_zz<0b000, "zip1", AArch64zip1>;
   defm ZIP2_ZZZ : sve_int_perm_bin_perm_zz<0b001, "zip2", AArch64zip2>;
   defm UZP1_ZZZ : sve_int_perm_bin_perm_zz<0b010, "uzp1", AArch64uzp1>;
   defm UZP2_ZZZ : sve_int_perm_bin_perm_zz<0b011, "uzp2", AArch64uzp2>;
   defm TRN1_ZZZ : sve_int_perm_bin_perm_zz<0b100, "trn1", AArch64trn1>;
   defm TRN2_ZZZ : sve_int_perm_bin_perm_zz<0b101, "trn2", AArch64trn2>;

llvm/test/CodeGen/AArch64/sve-intrinsics-adr.ll

+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -verify-machineinstrs < %s | FileCheck %s
+
+;
+; ADRB
+;
+
+define <vscale x 4 x i32> @adrb_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: adrb_i32:
+; CHECK: adr z0.s, [z0.s, z1.s]
+; CHECK-NEXT: ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> %a,
+                                                                <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @adrb_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: adrb_i64:
+; CHECK: adr z0.d, [z0.d, z1.d]
+; CHECK-NEXT: ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64> %a,
+                                                                <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+;
+; ADRH
+;
+
+define <vscale x 4 x i32> @adrh_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: adrh_i32:
+; CHECK: adr z0.s, [z0.s, z1.s, lsl #1]
+; CHECK-NEXT: ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> %a,
+                                                                <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @adrh_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: adrh_i64:
+; CHECK: adr z0.d, [z0.d, z1.d, lsl #1]
+; CHECK-NEXT: ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> %a,
+                                                                <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+;
+; ADRW
+;
+
+define <vscale x 4 x i32> @adrw_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: adrw_i32:
+; CHECK: adr z0.s, [z0.s, z1.s, lsl #2]
+; CHECK-NEXT: ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> %a,
+                                                                <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @adrw_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: adrw_i64:
+; CHECK: adr z0.d, [z0.d, z1.d, lsl #2]
+; CHECK-NEXT: ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> %a,
+                                                                <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+;
+; ADRD
+;
+
+define <vscale x 4 x i32> @adrd_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: adrd_i32:
+; CHECK: adr z0.s, [z0.s, z1.s, lsl #3]
+; CHECK-NEXT: ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> %a,
+                                                                <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @adrd_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: adrd_i64:
+; CHECK: adr z0.d, [z0.d, z1.d, lsl #3]
+; CHECK-NEXT: ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> %a,
+                                                                <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)