llvm · jthackray · Dec 16, 2024 · Dec 13, 2024 · Dec 16, 2024 · Dec 16, 2024
diff --git a/clang/include/clang/Basic/arm_sme.td b/clang/include/clang/Basic/arm_sme.td
@@ -749,6 +749,9 @@ let SMETargetGuard = "sme-f8f32" in {
   def SVDOT_LANE_FP8_ZA32_VG1x2 : Inst<"svdot_lane_za32[_mf8]_vg1x2_fpm", "vm2di>", "m", MergeNone, "aarch64_sme_fp8_fdot_lane_za32_vg1x2", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], [ImmCheck<3, ImmCheck0_3>]>;
   def SVDOT_LANE_FP8_ZA32_VG1x4 : Inst<"svdot_lane_za32[_mf8]_vg1x4_fpm", "vm4di>", "m", MergeNone, "aarch64_sme_fp8_fdot_lane_za32_vg1x4", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], [ImmCheck<3, ImmCheck0_3>]>;
 
+  def SVVDOTB_LANE_FP8_ZA32_VG1x4 : Inst<"svvdotb_lane_za32[_mf8]_vg1x4_fpm", "vm2di>", "m", MergeNone, "aarch64_sme_fp8_fvdotb_lane_za32_vg1x4", [IsOverloadNone, IsStreaming, IsInOutZA, SetsFPMR], [ImmCheck<3, ImmCheck0_3>]>;
+  def SVVDOTT_LANE_FP8_ZA32_VG1x4 : Inst<"svvdott_lane_za32[_mf8]_vg1x4_fpm", "vm2di>", "m", MergeNone, "aarch64_sme_fp8_fvdott_lane_za32_vg1x4", [IsOverloadNone, IsStreaming, IsInOutZA, SetsFPMR], [ImmCheck<3, ImmCheck0_3>]>;
+
   def SVDOT_SINGLE_FP8_ZA32_VG1x2 : Inst<"svdot[_single]_za32[_mf8]_vg1x2_fpm", "vm2d>", "m", MergeNone, "aarch64_sme_fp8_fdot_single_za32_vg1x2", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], []>;
   def SVDOT_SINGLE_FP8_ZA32_VG1x4 : Inst<"svdot[_single]_za32[_mf8]_vg1x4_fpm", "vm4d>", "m", MergeNone, "aarch64_sme_fp8_fdot_single_za32_vg1x4", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], []>;
 
@@ -760,6 +763,8 @@ let SMETargetGuard = "sme-f8f16" in {
   def SVDOT_LANE_FP8_ZA16_VG1x2 : Inst<"svdot_lane_za16[_mf8]_vg1x2_fpm", "vm2di>", "m", MergeNone, "aarch64_sme_fp8_fdot_lane_za16_vg1x2", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], [ImmCheck<3, ImmCheck0_7>]>;
   def SVDOT_LANE_FP8_ZA16_VG1x4 : Inst<"svdot_lane_za16[_mf8]_vg1x4_fpm", "vm4di>", "m", MergeNone, "aarch64_sme_fp8_fdot_lane_za16_vg1x4", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], [ImmCheck<3, ImmCheck0_7>]>;
 
+  def SVVDOT_LANE_FP8_ZA16_VG1x2 : Inst<"svvdot_lane_za16[_mf8]_vg1x2_fpm", "vm2di>", "m", MergeNone, "aarch64_sme_fp8_fvdot_lane_za16_vg1x2", [IsOverloadNone, IsStreaming, IsInOutZA, SetsFPMR], [ImmCheck<3, ImmCheck0_7>]>;
+
   def SVDOT_SINGLE_FP8_ZA16_VG1x2 : Inst<"svdot[_single]_za16[_mf8]_vg1x2_fpm", "vm2d>", "m", MergeNone, "aarch64_sme_fp8_fdot_single_za16_vg1x2", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], []>;
   def SVDOT_SINGLE_FP8_ZA16_VG1x4 : Inst<"svdot[_single]_za16[_mf8]_vg1x4_fpm", "vm4d>", "m", MergeNone, "aarch64_sme_fp8_fdot_single_za16_vg1x4", [IsStreaming, IsInOutZA, SetsFPMR, IsOverloadNone], []>;
 

diff --git a/clang/test/CodeGen/AArch64/sme2-intrinsics/acle_sme2_fp8_fvdot.c b/clang/test/CodeGen/AArch64/sme2-intrinsics/acle_sme2_fp8_fvdot.c
@@ -0,0 +1,80 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 4
+
+// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +sme-f8f16 -target-feature +sme-f8f32 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes mem2reg,instcombine,tailcallelim | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +sme-f8f16 -target-feature +sme-f8f32 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +sme-f8f16 -target-feature +sme-f8f32 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes mem2reg,instcombine,tailcallelim | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +sme-f8f16 -target-feature +sme-f8f32 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +sme-f8f16 -target-feature +sme-f8f32 -target-feature -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
+
+// REQUIRES: aarch64-registered-target
+
+#include <arm_sme.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+#define SVE_ACLE_FUNC(A1, A2_UNUSED, A3) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1, A2, A3) A1##A2##A3
+#endif
+
+// CHECK-LABEL: define dso_local void @test_svvdot_lane_za16_mf8_vg1x2(
+// CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdot.lane.za16.vg1x2(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 7)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: define dso_local void @_Z31test_svvdot_lane_za16_mf8_vg1x2j13svmfloat8x2_tu13__SVMfloat8_tm(
+// CPP-CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0:[0-9]+]] {
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdot.lane.za16.vg1x2(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 7)
+// CPP-CHECK-NEXT:    ret void
+//
+void test_svvdot_lane_za16_mf8_vg1x2(uint32_t slice, svmfloat8x2_t zn,
+                                    svmfloat8_t zm,
+                                    fpm_t fpmr) __arm_streaming
+    __arm_inout("za") {
+  SVE_ACLE_FUNC(svvdot_lane_za16, _mf8, _vg1x2_fpm)(slice, zn, zm, 7, fpmr);
+}
+
+// CHECK-LABEL: define dso_local void @test_svvdotb_lane_za32_mf8_vg1x4(
+// CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdotb.lane.za32.vg1x4(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: define dso_local void @_Z32test_svvdotb_lane_za32_mf8_vg1x4j13svmfloat8x2_tu13__SVMfloat8_tm(
+// CPP-CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdotb.lane.za32.vg1x4(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
+void test_svvdotb_lane_za32_mf8_vg1x4(uint32_t slice, svmfloat8x2_t zn,
+                                     svmfloat8_t zm,
+                                     fpm_t fpmr) __arm_streaming
+    __arm_inout("za") {
+  SVE_ACLE_FUNC(svvdotb_lane_za32, _mf8, _vg1x4_fpm)(slice, zn, zm, 3, fpmr);
+}
+
+// CHECK-LABEL: define dso_local void @test_svvdott_lane_za32_mf8_vg1x4(
+// CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdott.lane.za32.vg1x4(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: define dso_local void @_Z32test_svvdott_lane_za32_mf8_vg1x4j13svmfloat8x2_tu13__SVMfloat8_tm(
+// CPP-CHECK-SAME: i32 noundef [[SLICE:%.*]], <vscale x 16 x i8> [[ZN_COERCE0:%.*]], <vscale x 16 x i8> [[ZN_COERCE1:%.*]], <vscale x 16 x i8> [[ZM:%.*]], i64 noundef [[FPMR:%.*]]) #[[ATTR0]] {
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.set.fpmr(i64 [[FPMR]])
+// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fp8.fvdott.lane.za32.vg1x4(i32 [[SLICE]], <vscale x 16 x i8> [[ZN_COERCE0]], <vscale x 16 x i8> [[ZN_COERCE1]], <vscale x 16 x i8> [[ZM]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
+void test_svvdott_lane_za32_mf8_vg1x4(uint32_t slice, svmfloat8x2_t zn,
+                                     svmfloat8_t zm,
+                                     fpm_t fpmr) __arm_streaming
+    __arm_inout("za") {
+  SVE_ACLE_FUNC(svvdott_lane_za32, _mf8, _vg1x4_fpm)(slice, zn, zm, 3, fpmr);
+}
diff --git a/clang/test/Sema/aarch64-sme2-intrinsics/acle_sme2_fp8_fvdot.c b/clang/test/Sema/aarch64-sme2-intrinsics/acle_sme2_fp8_fvdot.c
@@ -0,0 +1,32 @@
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -target-feature +sme2 -verify -emit-llvm -o - %s
+
+// REQUIRES: aarch64-registered-target
+
+#include <arm_sme.h>
+
+void test_features(uint32_t slice, fpm_t fpmr, svmfloat8x2_t zn,
+                   svmfloat8_t zm) __arm_streaming __arm_inout("za") {
+// expected-error@+1 {{'svvdot_lane_za16_mf8_vg1x2_fpm' needs target feature sme,sme-f8f16}}
+  svvdot_lane_za16_mf8_vg1x2_fpm(slice, zn, zm, 7, fpmr);
+// expected-error@+1 {{'svvdotb_lane_za32_mf8_vg1x4_fpm' needs target feature sme,sme-f8f32}}
+  svvdotb_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, 3, fpmr);
+// expected-error@+1 {{'svvdott_lane_za32_mf8_vg1x4_fpm' needs target feature sme,sme-f8f32}}
+  svvdott_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, 3, fpmr);
+}
+
+void test_imm(uint32_t slice, fpm_t fpmr, svmfloat8x2_t zn,
+              svmfloat8_t zm) __arm_streaming __arm_inout("za") {
+// expected-error@+1 {{argument value 18446744073709551615 is outside the valid range [0, 7]}}
+  svvdot_lane_za16_mf8_vg1x2_fpm(slice, zn, zm, -1, fpmr);
+// expected-error@+1 {{argument value 18446744073709551615 is outside the valid range [0, 3]}}
+  svvdotb_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, -1, fpmr);
+// expected-error@+1 {{argument value 18446744073709551615 is outside the valid range [0, 3]}}
+  svvdott_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, -1, fpmr);
+
+// expected-error@+1{{argument value 8 is outside the valid range [0, 7]}}
+  svvdot_lane_za16_mf8_vg1x2_fpm(slice, zn, zm, 8, fpmr);
+// expected-error@+1{{argument value 4 is outside the valid range [0, 3]}}
+  svvdotb_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, 4, fpmr);
+// expected-error@+1{{argument value 4 is outside the valid range [0, 3]}}
+  svvdott_lane_za32_mf8_vg1x4_fpm(slice, zn, zm, 4, fpmr);
+}
diff --git a/llvm/include/llvm/IR/IntrinsicsAArch64.td b/llvm/include/llvm/IR/IntrinsicsAArch64.td
@@ -3904,6 +3904,11 @@ class SME2_FP8_FDOT_MULTI_VG1x4 :
   def int_aarch64_sme_fp8_fdot_lane_za32_vg1x2 : SME2_FP8_FDOT_LANE_VG1x2;
   def int_aarch64_sme_fp8_fdot_lane_za32_vg1x4 : SME2_FP8_FDOT_LANE_VG1x4;
 
+  def int_aarch64_sme_fp8_fvdot_lane_za16_vg1x2 : SME2_FP8_FDOT_LANE_VG1x2;
+
+  def int_aarch64_sme_fp8_fvdotb_lane_za32_vg1x4 : SME2_FP8_FDOT_LANE_VG1x2;
+  def int_aarch64_sme_fp8_fvdott_lane_za32_vg1x4 : SME2_FP8_FDOT_LANE_VG1x2;
+
   def int_aarch64_sme_fp8_fdot_single_za16_vg1x2 : SME2_FP8_FDOT_SINGLE_VG1x2;
   def int_aarch64_sme_fp8_fdot_single_za16_vg1x4 : SME2_FP8_FDOT_SINGLE_VG1x4;
 

diff --git a/llvm/lib/Target/AArch64/AArch64SMEInstrInfo.td b/llvm/lib/Target/AArch64/AArch64SMEInstrInfo.td
@@ -985,8 +985,8 @@ def LUTI4_S_4ZZT2Z  : sme2_luti4_vector_vg4_strided<0b00, 0b00, "luti4">;
 } //[HasSME2p1, HasSME_LUTv2]
 
 let Predicates = [HasSMEF8F16] in {
-defm FVDOT_VG2_M2ZZI_BtoH : sme2p1_multi_vec_array_vg2_index_f8f16<"fvdot", 0b11, 0b110, ZZ_b_mul_r, ZPR4b8>;
-defm FDOT_VG2_M2ZZI_BtoH  : sme2_fp8_fdot_index_za16_vg1x2<"fdot", int_aarch64_sme_fp8_fdot_lane_za16_vg1x2>;
+defm FVDOT_VG2_M2ZZI_BtoH : sme2_fp8_fdot_index_za16_vg1x2<"fvdot", 0b110, int_aarch64_sme_fp8_fvdot_lane_za16_vg1x2>;
+defm FDOT_VG2_M2ZZI_BtoH  : sme2_fp8_fdot_index_za16_vg1x2<"fdot",  0b010, int_aarch64_sme_fp8_fdot_lane_za16_vg1x2>;
 defm FDOT_VG4_M4ZZI_BtoH  : sme2_fp8_fdot_index_za16_vg1x4<"fdot", int_aarch64_sme_fp8_fdot_lane_za16_vg1x4>;
 
 defm FDOT_VG2_M2ZZ_BtoH   : sme2_fp8_fdot_single_vg1x2<"fdot", 0b0010001, MatrixOp16, int_aarch64_sme_fp8_fdot_single_za16_vg1x2>;
@@ -1016,8 +1016,8 @@ defm FDOT_VG4_M4ZZ_BtoS  : sme2_fp8_fdot_single_vg1x4<"fdot", 0b0110011, MatrixO
 defm FDOT_VG2_M2Z2Z_BtoS : sme2_fp8_fdot_multi_vg1x2 <"fdot", 0b0100110, MatrixOp32, int_aarch64_sme_fp8_fdot_multi_za32_vg1x2>;
 defm FDOT_VG4_M4Z4Z_BtoS : sme2_fp8_fdot_multi_vg1x4 <"fdot", 0b0100110, MatrixOp32, int_aarch64_sme_fp8_fdot_multi_za32_vg1x4>;
 
-def FVDOTB_VG4_M2ZZI_BtoS : sme2_fp8_multi_vec_array_vg4_index<"fvdotb", 0b0>;
-def FVDOTT_VG4_M2ZZI_BtoS : sme2_fp8_multi_vec_array_vg4_index<"fvdott", 0b1>;
+defm FVDOTB_VG4_M2ZZI_BtoS : sme2_fp8_fdotv_index_za32_vg1x4<"fvdotb", 0b0, int_aarch64_sme_fp8_fvdotb_lane_za32_vg1x4>;
+defm FVDOTT_VG4_M2ZZI_BtoS : sme2_fp8_fdotv_index_za32_vg1x4<"fvdott", 0b1, int_aarch64_sme_fp8_fvdott_lane_za32_vg1x4>;
 
 defm FMLALL_MZZI_BtoS      : sme2_mla_ll_array_index_32b<"fmlall",     0b01, 0b000, null_frag>;
 defm FMLALL_VG2_M2ZZI_BtoS : sme2_mla_ll_array_vg2_index_32b<"fmlall", 0b10, 0b100, null_frag>;

diff --git a/llvm/lib/Target/AArch64/SMEInstrFormats.td b/llvm/lib/Target/AArch64/SMEInstrFormats.td
@@ -5797,9 +5797,9 @@ multiclass sme2_fmop4a_fp8_fp16_2way<string mnemonic> {
 
 // FP8 SME FDOT instructions
 
-multiclass sme2_fp8_fdot_index_za16_vg1x2<string mnemonic,
+multiclass sme2_fp8_fdot_index_za16_vg1x2<string mnemonic, bits<3> op,
                                           SDPatternOperator intrinsic> {
-  def NAME : sme2_multi_vec_array_vg2_index<0b11, {0b0,?,?,0b10,?}, MatrixOp16,
+  def NAME : sme2_multi_vec_array_vg2_index<0b11, {op{2},?,?,op{1-0},?}, MatrixOp16,
                                             ZZ_b_mul_r, ZPR4b8,
                                             VectorIndexH32b_timm, mnemonic>,
                                             SMEPseudo2Instr<NAME, 1>{
@@ -5883,6 +5883,15 @@ multiclass sme2_fp8_fdot_index_za32_vg1x4<string mnemonic,
   def : SME2_ZA_TwoOp_VG4_Multi_Index_Pat<NAME, intrinsic, sme_elm_idx0_7, ZPR4b8, nxv16i8, VectorIndexS32b_timm, tileslice16>;
 }
 
+multiclass sme2_fp8_fdotv_index_za32_vg1x4<string mnemonic, bit T, SDPatternOperator intrinsic> {
+  def NAME : sme2_fp8_multi_vec_array_vg4_index<mnemonic, T>,
+                                            SMEPseudo2Instr<NAME, 1>;
+
+  def _PSEUDO : sme2_za_array_2op_multi_index_pseudo<NAME, sme_elm_idx0_7, ZZ_b_mul_r, ZPR4b8, VectorIndexS32b_timm, SMEMatrixArray>;
+
+  def : SME2_ZA_TwoOp_VG2_Multi_Index_Pat<NAME, intrinsic, sme_elm_idx0_7, ZPR4b8, nxv16i8, VectorIndexS32b_timm, tileslice16>;
+}
+
 multiclass sme2_fp8_fdot_single_vg1x2<string mnemonic, bits<7> op,
                                       MatrixOperand matrix_op,
                                       SDPatternOperator intrinsic> {

diff --git a/llvm/test/CodeGen/AArch64/sme2-intrinsics-fp8-fvdot.ll b/llvm/test/CodeGen/AArch64/sme2-intrinsics-fp8-fvdot.ll
@@ -0,0 +1,50 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --filter-out "// kill:" --version 4
+; RUN: llc -force-streaming < %s | FileCheck %s
+target triple = "aarch64-linux"
+
+define void @test_fvdot16_1x2_indexed(i32 %slice.0,
+; CHECK-LABEL: test_fvdot16_1x2_indexed:
+; CHECK:  // %bb.0:
+; CHECK:    mov w8, w0
+; CHECK:    fvdot za.h[w8, 7, vgx2], { z0.b, z1.b }, z2.b[3]
+; CHECK:    ret
+                                     <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                     <vscale x 16 x i8> %zm) #0 {
+  %slice = add i32 %slice.0, 7
+  call void @llvm.aarch64.sme.fp8.fvdot.lane.za16.vg1x2(i32 %slice,
+                                                       <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                                       <vscale x 16 x i8> %zm, i32 3)
+  ret void
+}
+
+define void @test_fvdot32_bottom_1x4_indexed(i32 %slice.0,
+; CHECK-LABEL: test_fvdot32_bottom_1x4_indexed:
+; CHECK:  // %bb.0:
+; CHECK:    mov w8, w0
+; CHECK:    fvdotb za.s[w8, 7, vgx4], { z0.b, z1.b }, z2.b[3]
+; CHECK:    ret
+                                             <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                             <vscale x 16 x i8> %zm) #0 {
+  %slice = add i32 %slice.0, 7
+  call void @llvm.aarch64.sme.fp8.fvdotb.lane.za32.vg1x4(i32 %slice,
+                                                         <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                                         <vscale x 16 x i8> %zm, i32 3)
+  ret void
+}
+
+define void @test_fvdot32_top_1x4_indexed(i32 %slice.0,
+; CHECK-LABEL: test_fvdot32_top_1x4_indexed:
+; CHECK:  // %bb.0:
+; CHECK:    mov w8, w0
+; CHECK:    fvdott za.s[w8, 7, vgx4], { z0.b, z1.b }, z2.b[3]
+; CHECK:    ret
+                                          <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                          <vscale x 16 x i8> %zm) #0 {
+  %slice = add i32 %slice.0, 7
+  call void @llvm.aarch64.sme.fp8.fvdott.lane.za32.vg1x4(i32 %slice,
+                                                         <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2,
+                                                         <vscale x 16 x i8> %zm, i32 3)
+  ret void
+}
+
+attributes #0 = { "target-features" = "+sme-f8f32,+sme-f8f16" }