[PowerPC] Add intrinsics for MMA

[saghir]

Summary

This patch adds support for MMA intrinsics.

Authored by: Baptiste Saleil

[lei]

This comment was removed by lei.

[bsaleil]

LGTM

[amyk]

Also LGTM.

llvm/include/llvm/IR/IntrinsicsPowerPC.td

let TargetPrefix = "ppc" in {  // All PPC intrinsics start with "llvm.ppc.".
   class PowerPC_VSX_Intrinsic<string GCCIntSuffix, list<LLVMType> ret_types,
                               list<LLVMType> param_types,
                               list<IntrinsicProperty> properties>
     : GCCBuiltin<!strconcat("__builtin_vsx_", GCCIntSuffix)>,
       Intrinsic<ret_types, param_types, properties>;
 }
 
 //===----------------------------------------------------------------------===//
+// PowerPC MMA Intrinsic Multi Class Definitions.
+//
+
+multiclass PowerPC_MMA_ACC_Intrinsic<list<LLVMType> args> {
+  def NAME: Intrinsic<[llvm_v512i1_ty], args, [IntrNoMem]>;
+  def pp : Intrinsic<[llvm_v512i1_ty], !listconcat([llvm_v512i1_ty], args),
+                     [IntrNoMem]>;
+  def pn : Intrinsic<[llvm_v512i1_ty], !listconcat([llvm_v512i1_ty], args),
+                     [IntrNoMem]>;
+  def np : Intrinsic<[llvm_v512i1_ty], !listconcat([llvm_v512i1_ty], args),
+                     [IntrNoMem]>;
+  def nn : Intrinsic<[llvm_v512i1_ty], !listconcat([llvm_v512i1_ty], args),
+                     [IntrNoMem]>;
+}
+
+multiclass PowerPC_MMA_ACC_PP_Intrinsic<list<LLVMType> args> {
+  def NAME: Intrinsic<[llvm_v512i1_ty], args, [IntrNoMem]>;
+  def pp : Intrinsic<[llvm_v512i1_ty], !listconcat([llvm_v512i1_ty], args),
+                     [IntrNoMem]>;
+}
+
+//===----------------------------------------------------------------------===//
 // PowerPC Altivec Intrinsic Class Definitions.
 //
 
 /// PowerPC_Vec_FF_Intrinsic - A PowerPC intrinsic that takes one v4f32
 /// vector and returns one.  These intrinsics have no side effects.
 class PowerPC_Vec_FF_Intrinsic<string GCCIntSuffix>
   : PowerPC_Vec_Intrinsic<GCCIntSuffix,
                           [llvm_v4f32_ty], [llvm_v4f32_ty], [IntrNoMem]>;

 def int_ppc_ttest : GCCBuiltin<"__builtin_ttest">,
       Intrinsic<[llvm_i64_ty], [], []>;
 
 def int_ppc_cfence : Intrinsic<[], [llvm_anyint_ty], []>;
 
 // PowerPC set FPSCR Intrinsic Definitions.
 def int_ppc_setrnd : GCCBuiltin<"__builtin_setrnd">,
       Intrinsic<[llvm_double_ty], [llvm_i32_ty], []>;
-
 }
 
 let TargetPrefix = "ppc" in {
   def int_ppc_mma_assemble_acc :
         Intrinsic<[llvm_v512i1_ty],
                   [llvm_v16i8_ty, llvm_v16i8_ty, llvm_v16i8_ty, llvm_v16i8_ty],
                   [IntrNoMem]>;
 
let TargetPrefix = "ppc" in {
   def int_ppc_mma_xxmtacc :
         Intrinsic<[llvm_v512i1_ty], [llvm_v512i1_ty], [IntrNoMem]>;
 
   def int_ppc_mma_xxmfacc :
         Intrinsic<[llvm_v512i1_ty], [llvm_v512i1_ty], [IntrNoMem]>;
 
   def int_ppc_mma_xxsetaccz :
         Intrinsic<[llvm_v512i1_ty], [], [IntrNoMem]>;
-}
 
+  // MMA Reduced-Precision: Outer Product Intrinsic Definitions.
+  defm int_ppc_mma_xvi4ger8 :
+        PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvi4ger8 :
+        PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                      llvm_i32_ty, llvm_i32_ty]>;
+
+  defm int_ppc_mma_xvi8ger4 :
+       PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvi8ger4 :
+       PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                     llvm_i32_ty, llvm_i32_ty]>;
+
+  defm int_ppc_mma_xvi16ger2s :
+       PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvi16ger2s :
+       PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                     llvm_i32_ty, llvm_i32_ty]>;
+
+  defm int_ppc_mma_xvf16ger2 :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvf16ger2 :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                  llvm_i32_ty, llvm_i32_ty]>;
+  defm int_ppc_mma_xvf32ger :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvf32ger :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                  llvm_i32_ty]>;
+  defm int_ppc_mma_xvf64ger :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v256i1_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvf64ger :
+       PowerPC_MMA_ACC_Intrinsic<[llvm_v256i1_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                  llvm_i32_ty]>;
+
+  // MMA Reduced-Precision: bfloat16 Outer Product Intrinsic Definitions.
+  defm int_ppc_mma_xvbf16ger2 :
+         PowerPC_MMA_ACC_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvbf16ger2 :
+         PowerPC_MMA_ACC_Intrinsic<
+           [llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty, llvm_i32_ty, llvm_i32_ty]>;
+
+  // MMA Reduced-Precision: Missing Integer-based Outer Product Operations.
+  defm int_ppc_mma_xvi16ger2 :
+         PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty]>;
+  defm int_ppc_mma_pmxvi16ger2 :
+         PowerPC_MMA_ACC_PP_Intrinsic<[llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                                       llvm_i32_ty, llvm_i32_ty]>;
+  def int_ppc_mma_xvi8ger4spp :
+        Intrinsic<[llvm_v512i1_ty],
+                  [llvm_v512i1_ty, llvm_v16i8_ty, llvm_v16i8_ty], [IntrNoMem]>;
+  def int_ppc_mma_pmxvi8ger4spp :
+        Intrinsic<[llvm_v512i1_ty],
+                  [llvm_v512i1_ty, llvm_v16i8_ty, llvm_v16i8_ty, llvm_i32_ty,
+                   llvm_i32_ty, llvm_i32_ty], [IntrNoMem]>;
+}

llvm/lib/Target/PowerPC/PPCInstrPrefix.td

+// Mask immediates for MMA instructions (2, 4 and 8 bits).
+def Msk2Imm : ImmLeaf<i32, [{ return isUInt<2>(Imm); }]>;
+def Msk4Imm : ImmLeaf<i32, [{ return isUInt<4>(Imm); }]>;
+def Msk8Imm : ImmLeaf<i32, [{ return isUInt<8>(Imm); }]>;
+
 //===----------------------------------------------------------------------===//
 // PowerPC ISA 3.1 specific type constraints.
 //
 
 def SDT_PPCSplat32 : SDTypeProfile<1, 3, [ SDTCisVT<0, v2i64>,
   SDTCisVec<1>, SDTCisInt<2>, SDTCisInt<3>
 ]>;
 def SDT_PPCAccBuild : SDTypeProfile<1, 4, [
defm XVF32GER : ACC_NEG_UM_M44_XOM84C<59, 26, (ins vsrc:$XA, vsrc:$XB),
                                       "xvf32ger", "$AT, $XA, $XB">;
 
 // XVF64GER, XVF64GERPP, XVF64GERPN, XVF64GERNP, XVF64GERNN
 // PMXVF64GER, PMXVF64GERPP, PMXVF64GERPN, PMXVF64GERNP, PMXVF64GERNN
 defm XVF64GER : ACC_NEG_UM_M42_XOM84C<59, 58, (ins vsrpevenrc:$XA, vsrc:$XB),
                                       "xvf64ger", "$AT, $XA, $XB">;
 //------------------------------------------------------------------------------
 
+// MMA Intrinsics
+let Predicates = [MMA] in {
+  def : Pat<(v512i1 (int_ppc_mma_xvi4ger8 v16i8:$XA, v16i8:$XB)),
+            (XVI4GER8 RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi4ger8pp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVI4GER8PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_xvi8ger4 v16i8:$XA, v16i8:$XB)),
+            (XVI8GER4 RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi8ger4pp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVI8GER4PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_xvi16ger2s v16i8:$XA, v16i8:$XB)),
+            (XVI16GER2S RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi16ger2spp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVI16GER2SPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_xvf16ger2 v16i8:$XA, v16i8:$XB)),
+            (XVF16GER2 RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf16ger2pn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF16GER2PN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf16ger2np v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF16GER2NP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf16ger2nn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF16GER2NN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_xvf32ger v16i8:$XA, v16i8:$XB)),
+            (XVF32GER RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf32gerpp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF32GERPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf32gerpn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF32GERPN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf32gernp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF32GERNP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf32gernn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVF32GERNN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf64ger v256i1:$XA, v16i8:$XB)),
+            (XVF64GER $XA, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf64gerpp v512i1:$ATi, v256i1:$XA, v16i8:$XB)),
+            (XVF64GERPP $ATi, $XA, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf64gerpn v512i1:$ATi, v256i1:$XA, v16i8:$XB)),
+            (XVF64GERPN $ATi, $XA, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf64gernp v512i1:$ATi, v256i1:$XA, v16i8:$XB)),
+            (XVF64GERNP $ATi, $XA, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvf64gernn v512i1:$ATi, v256i1:$XA, v16i8:$XB)),
+            (XVF64GERNN $ATi, $XA, RCCp.BToVSRC)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_xvbf16ger2 v16i8:$XA, v16i8:$XB)),
+            (XVBF16GER2 RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvbf16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVBF16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvbf16ger2pn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVBF16GER2PN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvbf16ger2np v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVBF16GER2NP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvbf16ger2nn v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVBF16GER2NN $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi16ger2 v16i8:$XA, v16i8:$XB)),
+            (XVI16GER2 RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVI16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+  def : Pat<(v512i1 (int_ppc_mma_xvi8ger4spp v512i1:$ATi, v16i8:$XA, v16i8:$XB)),
+            (XVI8GER4SPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC)>;
+}
+
+// MMA Intrinsics
+let Predicates = [MMA, PrefixInstrs] in {
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi4ger8 v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                            Msk4Imm:$YMSK, Msk8Imm:$PMSK)),
+            (PMXVI4GER8 RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                        Msk4Imm:$YMSK, Msk8Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi4ger8pp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                              Msk8Imm:$PMSK)),
+            (PMXVI4GER8PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK, Msk8Imm:$PMSK)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi8ger4 v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                            Msk4Imm:$YMSK, Msk4Imm:$PMSK)),
+            (PMXVI8GER4 RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                        Msk4Imm:$YMSK, Msk4Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi8ger4pp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                              Msk4Imm:$PMSK)),
+            (PMXVI8GER4PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK, Msk4Imm:$PMSK)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi16ger2s v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                              Msk4Imm:$YMSK, Msk2Imm:$PMSK)),
+            (PMXVI16GER2S RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi16ger2spp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                                Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                                Msk2Imm:$PMSK)),
+            (PMXVI16GER2SPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                            Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2 v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                             Msk4Imm:$YMSK, Msk2Imm:$PMSK)),
+            (PMXVF16GER2 RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                         Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2pn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2PN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2np v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2NP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2pn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2PN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2np v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2NP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf16ger2nn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVF16GER2NN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf32ger v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                            Msk4Imm:$YMSK)),
+            (PMXVF32GER RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                        Msk4Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf32gerpp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK)),
+            (PMXVF32GERPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf32gerpn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK)),
+            (PMXVF32GERPN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf32gernp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK)),
+            (PMXVF32GERNP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf32gernn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk4Imm:$YMSK)),
+            (PMXVF32GERNN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf64ger v256i1:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                            Msk2Imm:$YMSK)),
+            (PMXVF64GER $XA, RCCp.BToVSRC, Msk4Imm:$XMSK, Msk2Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf64gerpp v512i1:$ATi, v256i1:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk2Imm:$YMSK)),
+            (PMXVF64GERPP $ATi, $XA, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk2Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf64gerpn v512i1:$ATi, v256i1:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk2Imm:$YMSK)),
+            (PMXVF64GERPN $ATi, $XA, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk2Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf64gernp v512i1:$ATi, v256i1:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk2Imm:$YMSK)),
+            (PMXVF64GERNP $ATi, $XA, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk2Imm:$YMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvf64gernn v512i1:$ATi, v256i1:$XA, v16i8:$XB,
+                                              Msk4Imm:$XMSK, Msk2Imm:$YMSK)),
+            (PMXVF64GERNN $ATi, $XA, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk2Imm:$YMSK)>;
+
+  def : Pat<(v512i1 (int_ppc_mma_pmxvbf16ger2 v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                              Msk4Imm:$YMSK, Msk2Imm:$PMSK)),
+            (PMXVBF16GER2 RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                          Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvbf16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                                Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                                Msk2Imm:$PMSK)),
+            (PMXVBF16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                            Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvbf16ger2pn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                                Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                                Msk2Imm:$PMSK)),
+            (PMXVBF16GER2PN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                            Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvbf16ger2np v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                                Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                                Msk2Imm:$PMSK)),
+            (PMXVBF16GER2NP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                            Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvbf16ger2nn v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                                Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                                Msk2Imm:$PMSK)),
+            (PMXVBF16GER2NN $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                            Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi16ger2 v16i8:$XA, v16i8:$XB, Msk4Imm:$XMSK,
+                                             Msk4Imm:$YMSK, Msk2Imm:$PMSK)),
+            (PMXVI16GER2 RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                         Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi8ger4spp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVI8GER4SPP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+  def : Pat<(v512i1 (int_ppc_mma_pmxvi16ger2pp v512i1:$ATi, v16i8:$XA, v16i8:$XB,
+                                               Msk4Imm:$XMSK, Msk4Imm:$YMSK,
+                                               Msk2Imm:$PMSK)),
+            (PMXVI16GER2PP $ATi, RCCp.AToVSRC, RCCp.BToVSRC, Msk4Imm:$XMSK,
+                           Msk4Imm:$YMSK, Msk2Imm:$PMSK)>;
+}
+
 def Concats {
   dag VecsToVecPair0 =
     (v256i1 (INSERT_SUBREG
       (INSERT_SUBREG (IMPLICIT_DEF), $vs0, sub_vsx1),
       $vs1, sub_vsx0));
   dag VecsToVecPair1 =
     (v256i1 (INSERT_SUBREG
       (INSERT_SUBREG (IMPLICIT_DEF), $vs2, sub_vsx1),

llvm/test/CodeGen/PowerPC/bfloat16-outer-product.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
 ; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
 ; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
 ; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
 ; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s --check-prefix=CHECK-BE
 
 ; Function Attrs: nofree nounwind writeonly
+define dso_local void @test50(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test50:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvbf16ger2 acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test50:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvbf16ger2 acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvbf16ger2(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvbf16ger2(<16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind writeonly
+define dso_local void @test51(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test51:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvbf16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test51:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvbf16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvbf16ger2(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvbf16ger2(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test52(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test52:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvbf16ger2pp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test52:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvbf16ger2pp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvbf16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvbf16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test53(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test53:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvbf16ger2pn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test53:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvbf16ger2pn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvbf16ger2pn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvbf16ger2pn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test54(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test54:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvbf16ger2np acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test54:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvbf16ger2np acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvbf16ger2np(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvbf16ger2np(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test55(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test55:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvbf16ger2nn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test55:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvbf16ger2nn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvbf16ger2nn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvbf16ger2nn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test56(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test56:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvbf16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test56:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvbf16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvbf16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvbf16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test57(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test57:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvbf16ger2pn acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test57:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvbf16ger2pn acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvbf16ger2pn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvbf16ger2pn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test58(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test58:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvbf16ger2np acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test58:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvbf16ger2np acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvbf16ger2np(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvbf16ger2np(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test59(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test59:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvbf16ger2nn acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test59:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvbf16ger2nn acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvbf16ger2nn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvbf16ger2nn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind writeonly
 define dso_local void @test60(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
 ; CHECK-LABEL: test60:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvcvspbf16 vs0, v2
 ; CHECK-NEXT:    stxv vs0, 0(r7)
 ; CHECK-NEXT:    blr
 ;
 ; CHECK-BE-LABEL: test60:
 ; CHECK-BE:       # %bb.0: # %entry
 ; CHECK-BE-NEXT:    xvcvspbf16 vs0, v2
 ; CHECK-BE-NEXT:    stxv vs0, 0(r7)
 ; CHECK-BE-NEXT:    blr
 entry:
   %0 = tail call <16 x i8> @llvm.ppc.vsx.xvcvspbf16(<16 x i8> %vc)
   %1 = bitcast i8* %resp to <16 x i8>*
   store <16 x i8> %0, <16 x i8>* %1, align 16
   ret void
 }
+
 ; Function Attrs: nounwind readnone
 declare <16 x i8> @llvm.ppc.vsx.xvcvspbf16(<16 x i8>)
 
 ; Function Attrs: nofree nounwind writeonly
 define dso_local void @test61(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
 ; CHECK-LABEL: test61:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvcvbf16spn vs0, v2

llvm/test/CodeGen/PowerPC/mma-integer-based-outer-product.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
+; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
+; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s --check-prefix=CHECK-BE
+
+; Function Attrs: nofree nounwind writeonly
+define dso_local void @test1(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvi16ger2 acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test1:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvi16ger2 acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvi16ger2(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvi16ger2(<16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind writeonly
+define dso_local void @test2(i8* nocapture readnone %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvi16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test2:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvi16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvi16ger2(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvi16ger2(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test3(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi8ger4spp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test3:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi8ger4spp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi8ger4spp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvi8ger4spp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test4(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi16ger2pp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test4:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi16ger2pp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.xvi16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test5(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test5:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvi8ger4spp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test5:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvi8ger4spp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvi8ger4spp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvi8ger4spp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+; Function Attrs: nofree nounwind
+define dso_local void @test6(i8* nocapture readonly %vqp, i8* nocapture readnone %vpp, <16 x i8> %vc, i8* nocapture %resp) {
+; CHECK-LABEL: test6:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvi16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test6:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvi16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvi16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+; Function Attrs: nounwind readnone
+declare <512 x i1> @llvm.ppc.mma.pmxvi16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)

llvm/test/CodeGen/PowerPC/mma-outer-product.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
+; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names \
+; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s --check-prefix=CHECK-BE
+
+declare <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8>, <16 x i8>, <16 x i8>, <16 x i8>)
+declare <256 x i1> @llvm.ppc.mma.assemble.pair(<16 x i8>, <16 x i8>)
+define void @intrinsics1(<16 x i8> %vc1, <16 x i8> %vc2, <16 x i8> %vc3, <16 x i8> %vc4, i8* %ptr) {
+; CHECK-LABEL: intrinsics1:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    # kill: def $v5 killed $v5 killed $vsrp18 def $vsrp18
+; CHECK-NEXT:    vmr v1, v4
+; CHECK-NEXT:    vmr v4, v3
+; CHECK-NEXT:    ld r3, 96(r1)
+; CHECK-NEXT:    vmr v0, v2
+; CHECK-NEXT:    xxlor vs5, v2, v2
+; CHECK-NEXT:    xxlor vs4, v5, v5
+; CHECK-NEXT:    xxlor vs0, v0, v0
+; CHECK-NEXT:    xxlor vs1, v1, v1
+; CHECK-NEXT:    xxlor vs2, v4, v4
+; CHECK-NEXT:    xxlor vs3, v5, v5
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi4ger8pp acc0, v2, v3
+; CHECK-NEXT:    xvf16ger2pp acc0, v2, v1
+; CHECK-NEXT:    pmxvf32gerpn acc0, v3, v5, 0, 0
+; CHECK-NEXT:    pmxvf64gernp acc0, vsp4, v0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r3)
+; CHECK-NEXT:    stxv vs1, 32(r3)
+; CHECK-NEXT:    stxv vs2, 16(r3)
+; CHECK-NEXT:    stxvx vs3, 0, r3
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: intrinsics1:
+; CHECK-BE:       # %bb.0:
+; CHECK-BE-NEXT:    # kill: def $v5 killed $v5 killed $vsrp18 def $vsrp18
+; CHECK-BE-NEXT:    vmr v1, v4
+; CHECK-BE-NEXT:    vmr v4, v3
+; CHECK-BE-NEXT:    ld r3, 112(r1)
+; CHECK-BE-NEXT:    vmr v0, v2
+; CHECK-BE-NEXT:    xxlor vs5, v2, v2
+; CHECK-BE-NEXT:    xxlor vs4, v5, v5
+; CHECK-BE-NEXT:    xxlor vs0, v0, v0
+; CHECK-BE-NEXT:    xxlor vs1, v1, v1
+; CHECK-BE-NEXT:    xxlor vs2, v4, v4
+; CHECK-BE-NEXT:    xxlor vs3, v5, v5
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi4ger8pp acc0, v2, v3
+; CHECK-BE-NEXT:    xvf16ger2pp acc0, v2, v1
+; CHECK-BE-NEXT:    pmxvf32gerpn acc0, v3, v5, 0, 0
+; CHECK-BE-NEXT:    pmxvf64gernp acc0, vsp4, v0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r3)
+; CHECK-BE-NEXT:    stxvx vs0, 0, r3
+; CHECK-BE-NEXT:    stxv vs3, 48(r3)
+; CHECK-BE-NEXT:    stxv vs2, 32(r3)
+; CHECK-BE-NEXT:    blr
+  %1 = tail call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> %vc1, <16 x i8> %vc3, <16 x i8> %vc2, <16 x i8> %vc4)
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi4ger8pp(<512 x i1> %1, <16 x i8> %vc1, <16 x i8> %vc2)
+  %3 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1> %2, <16 x i8> %vc1, <16 x i8> %vc3)
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gerpn(<512 x i1> %3, <16 x i8> %vc2, <16 x i8> %vc4, i32 0, i32 0)
+  %5 = tail call <256 x i1> @llvm.ppc.mma.assemble.pair(<16 x i8> %vc4, <16 x i8> %vc1)
+  %6 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gernp(<512 x i1> %4, <256 x i1> %5, <16 x i8> %vc1, i32 0, i32 0)
+  %7 = bitcast i8* %ptr to <512 x i1>*
+  store <512 x i1> %6, <512 x i1>* %7, align 64
+  ret void
+}
+
+declare { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } @llvm.ppc.mma.disassemble.acc(<512 x i1>)
+define void @intrinsics2(<16 x i8>* %ptr1, <16 x i8>* %ptr2, <16 x i8>* %ptr3, <16 x i8>* %ptr4, i8* %ptr) {
+; CHECK-LABEL: intrinsics2:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lxv vs4, 0(r3)
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs6, 0(r5)
+; CHECK-NEXT:    lxv vs7, 0(r6)
+; CHECK-NEXT:    xxlor vs0, vs4, vs4
+; CHECK-NEXT:    xxlor vs9, vs4, vs4
+; CHECK-NEXT:    xxlor vs1, vs5, vs5
+; CHECK-NEXT:    xxlor vs2, vs6, vs6
+; CHECK-NEXT:    xxlor vs3, vs7, vs7
+; CHECK-NEXT:    xxlor vs8, vs7, vs7
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi8ger4pp acc0, vs4, vs5
+; CHECK-NEXT:    xvf16ger2pn acc0, vs4, vs6
+; CHECK-NEXT:    pmxvf32gernn acc0, vs5, vs7, 0, 0
+; CHECK-NEXT:    pmxvf64gernn acc0, vsp8, vs4, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs3, 0(r3)
+; CHECK-NEXT:    stxv vs2, 0(r4)
+; CHECK-NEXT:    stxv vs1, 0(r5)
+; CHECK-NEXT:    stxv vs0, 0(r6)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: intrinsics2:
+; CHECK-BE:       # %bb.0:
+; CHECK-BE-NEXT:    lxv vs4, 0(r3)
+; CHECK-BE-NEXT:    lxv vs5, 0(r4)
+; CHECK-BE-NEXT:    lxv vs6, 0(r5)
+; CHECK-BE-NEXT:    lxv vs7, 0(r6)
+; CHECK-BE-NEXT:    xxlor vs0, vs4, vs4
+; CHECK-BE-NEXT:    xxlor vs9, vs4, vs4
+; CHECK-BE-NEXT:    xxlor vs1, vs5, vs5
+; CHECK-BE-NEXT:    xxlor vs2, vs6, vs6
+; CHECK-BE-NEXT:    xxlor vs3, vs7, vs7
+; CHECK-BE-NEXT:    xxlor vs8, vs7, vs7
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi8ger4pp acc0, vs4, vs5
+; CHECK-BE-NEXT:    xvf16ger2pn acc0, vs4, vs6
+; CHECK-BE-NEXT:    pmxvf32gernn acc0, vs5, vs7, 0, 0
+; CHECK-BE-NEXT:    pmxvf64gernn acc0, vsp8, vs4, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs0, 0(r3)
+; CHECK-BE-NEXT:    stxv vs1, 0(r4)
+; CHECK-BE-NEXT:    stxv vs2, 0(r5)
+; CHECK-BE-NEXT:    stxv vs3, 0(r6)
+; CHECK-BE-NEXT:    blr
+  %vc1 = load <16 x i8>, <16 x i8>* %ptr1, align 16
+  %vc2 = load <16 x i8>, <16 x i8>* %ptr2, align 16
+  %vc3 = load <16 x i8>, <16 x i8>* %ptr3, align 16
+  %vc4 = load <16 x i8>, <16 x i8>* %ptr4, align 16
+  %1 = tail call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> %vc1, <16 x i8> %vc2, <16 x i8> %vc3, <16 x i8> %vc4)
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi8ger4pp(<512 x i1> %1, <16 x i8> %vc1, <16 x i8> %vc2)
+  %3 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pn(<512 x i1> %2, <16 x i8> %vc1, <16 x i8> %vc3)
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gernn(<512 x i1> %3, <16 x i8> %vc2, <16 x i8> %vc4, i32 0, i32 0)
+  %5 = tail call <256 x i1> @llvm.ppc.mma.assemble.pair(<16 x i8> %vc4, <16 x i8> %vc1)
+  %6 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gernn(<512 x i1> %4, <256 x i1> %5, <16 x i8> %vc1, i32 0, i32 0)
+  %7 = tail call { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } @llvm.ppc.mma.disassemble.acc(<512 x i1> %6)
+  %8 = extractvalue { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } %7, 0
+  %9 = extractvalue { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } %7, 1
+  %10 = extractvalue { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } %7, 2
+  %11 = extractvalue { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } %7, 3
+  %12 = bitcast i8* %ptr to <512 x i1>*
+  store <16 x i8> %8, <16 x i8>* %ptr1, align 16
+  store <16 x i8> %9, <16 x i8>* %ptr2, align 16
+  store <16 x i8> %10, <16 x i8>* %ptr3, align 16
+  store <16 x i8> %11, <16 x i8>* %ptr4, align 16
+  ret void
+}
+
+define void @test1(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvi4ger8 acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test1:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvi4ger8 acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvi4ger8(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi4ger8(<16 x i8>, <16 x i8>)
+
+define void @test2(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi4ger8pp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test2:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi4ger8pp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi4ger8pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi4ger8pp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test3(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvi4ger8 acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test3:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvi4ger8 acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvi4ger8(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi4ger8(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test4(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvi4ger8pp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test4:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvi4ger8pp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvi4ger8pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi4ger8pp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test5(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test5:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvi8ger4 acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test5:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvi8ger4 acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvi8ger4(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi8ger4(<16 x i8>, <16 x i8>)
+
+define void @test6(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test6:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi8ger4pp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test6:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi8ger4pp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi8ger4pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi8ger4pp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test7(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test7:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvi8ger4 acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test7:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvi8ger4 acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvi8ger4(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi8ger4(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test8(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvi8ger4pp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test8:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvi8ger4pp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvi8ger4pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi8ger4pp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test9(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test9:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvi16ger2s acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test9:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvi16ger2s acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvi16ger2s(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi16ger2s(<16 x i8>, <16 x i8>)
+
+define void @test10(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test10:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvi16ger2spp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test10:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvi16ger2spp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvi16ger2spp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvi16ger2spp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test11(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test11:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvi16ger2s acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test11:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvi16ger2s acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvi16ger2s(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi16ger2s(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test12(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test12:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvi16ger2spp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test12:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvi16ger2spp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvi16ger2spp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvi16ger2spp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test13(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test13:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvf16ger2 acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test13:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvf16ger2 acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf16ger2(<16 x i8>, <16 x i8>)
+
+define void @test14(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test14:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf16ger2pp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test14:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf16ger2pp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test15(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test15:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf16ger2pn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test15:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf16ger2pn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2pn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf16ger2pn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test16(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf16ger2np acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test16:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf16ger2np acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2np(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf16ger2np(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test17(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test17:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf16ger2nn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test17:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf16ger2nn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf16ger2nn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf16ger2nn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test18(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test18:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvf16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test18:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvf16ger2 acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvf16ger2(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf16ger2(<16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test19(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test19:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test19:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf16ger2pp acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf16ger2pp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf16ger2pp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test20(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test20:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf16ger2pn acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test20:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf16ger2pn acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf16ger2pn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf16ger2pn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test21(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test21:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf16ger2np acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test21:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf16ger2np acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf16ger2np(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf16ger2np(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test22(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test22:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf16ger2nn acc0, v2, v2, 0, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test22:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf16ger2nn acc0, v2, v2, 0, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf16ger2nn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf16ger2nn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32, i32)
+
+define void @test23(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test23:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xvf32ger acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test23:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvf32ger acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.xvf32ger(<16 x i8> %vc, <16 x i8> %vc)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf32ger(<16 x i8>, <16 x i8>)
+
+define void @test24(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test24:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf32gerpp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test24:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf32gerpp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf32gerpp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf32gerpp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test25(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test25:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf32gerpn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test25:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf32gerpn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf32gerpn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf32gerpn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test26(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test26:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf32gernp acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test26:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf32gernp acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf32gernp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf32gernp(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test27(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test27:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    xvf32gernn acc0, v2, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test27:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    xvf32gernn acc0, v2, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf32gernn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf32gernn(<512 x i1>, <16 x i8>, <16 x i8>)
+
+define void @test28(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test28:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pmxvf32ger acc0, v2, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test28:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    pmxvf32ger acc0, v2, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32ger(<16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0)
+  %1 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %0, <512 x i1>* %1, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf32ger(<16 x i8>, <16 x i8>, i32, i32)
+
+define void @test29(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test29:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf32gerpp acc0, v2, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test29:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf32gerpp acc0, v2, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gerpp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf32gerpp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32)
+
+define void @test30(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test30:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf32gerpn acc0, v2, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test30:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf32gerpn acc0, v2, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gerpn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf32gerpn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32)
+
+define void @test31(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test31:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf32gernp acc0, v2, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test31:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf32gernp acc0, v2, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gernp(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf32gernp(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32)
+
+define void @test32(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    pmxvf32gernn acc0, v2, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test32:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    pmxvf32gernn acc0, v2, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf32gernn(<512 x i1> %1, <16 x i8> %vc, <16 x i8> %vc, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf32gernn(<512 x i1>, <16 x i8>, <16 x i8>, i32, i32)
+
+define void @test33(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test33:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 0(r4)
+; CHECK-NEXT:    lxv vs0, 16(r4)
+; CHECK-NEXT:    xvf64ger acc1, vsp0, v2
+; CHECK-NEXT:    xxmfacc acc1
+; CHECK-NEXT:    stxv vs4, 48(r7)
+; CHECK-NEXT:    stxv vs5, 32(r7)
+; CHECK-NEXT:    stxv vs6, 16(r7)
+; CHECK-NEXT:    stxv vs7, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test33:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r4)
+; CHECK-BE-NEXT:    lxv vs0, 0(r4)
+; CHECK-BE-NEXT:    xvf64ger acc1, vsp0, v2
+; CHECK-BE-NEXT:    xxmfacc acc1
+; CHECK-BE-NEXT:    stxv vs5, 16(r7)
+; CHECK-BE-NEXT:    stxv vs4, 0(r7)
+; CHECK-BE-NEXT:    stxv vs7, 48(r7)
+; CHECK-BE-NEXT:    stxv vs6, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vpp to <256 x i1>*
+  %1 = load <256 x i1>, <256 x i1>* %0, align 32
+  %2 = tail call <512 x i1> @llvm.ppc.mma.xvf64ger(<256 x i1> %1, <16 x i8> %vc)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf64ger(<256 x i1>, <16 x i8>)
+
+define void @test34(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test34:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    xvf64gerpp acc0, vsp4, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test34:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    xvf64gerpp acc0, vsp4, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.xvf64gerpp(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf64gerpp(<512 x i1>, <256 x i1>, <16 x i8>)
+
+define void @test35(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test35:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    xvf64gerpn acc0, vsp4, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test35:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    xvf64gerpn acc0, vsp4, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.xvf64gerpn(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf64gerpn(<512 x i1>, <256 x i1>, <16 x i8>)
+
+define void @test36(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test36:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    xvf64gernp acc0, vsp4, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test36:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    xvf64gernp acc0, vsp4, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.xvf64gernp(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf64gernp(<512 x i1>, <256 x i1>, <16 x i8>)
+
+define void @test37(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test37:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    xvf64gernn acc0, vsp4, v2
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test37:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    xvf64gernn acc0, vsp4, v2
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.xvf64gernn(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.xvf64gernn(<512 x i1>, <256 x i1>, <16 x i8>)
+
+define void @test38(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test38:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 0(r4)
+; CHECK-NEXT:    lxv vs0, 16(r4)
+; CHECK-NEXT:    pmxvf64ger acc1, vsp0, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc1
+; CHECK-NEXT:    stxv vs4, 48(r7)
+; CHECK-NEXT:    stxv vs5, 32(r7)
+; CHECK-NEXT:    stxv vs6, 16(r7)
+; CHECK-NEXT:    stxv vs7, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test38:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r4)
+; CHECK-BE-NEXT:    lxv vs0, 0(r4)
+; CHECK-BE-NEXT:    pmxvf64ger acc1, vsp0, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc1
+; CHECK-BE-NEXT:    stxv vs5, 16(r7)
+; CHECK-BE-NEXT:    stxv vs4, 0(r7)
+; CHECK-BE-NEXT:    stxv vs7, 48(r7)
+; CHECK-BE-NEXT:    stxv vs6, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vpp to <256 x i1>*
+  %1 = load <256 x i1>, <256 x i1>* %0, align 32
+  %2 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64ger(<256 x i1> %1, <16 x i8> %vc, i32 0, i32 0)
+  %3 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %2, <512 x i1>* %3, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf64ger(<256 x i1>, <16 x i8>, i32, i32)
+
+define void @test39(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test39:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    pmxvf64gerpp acc0, vsp4, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test39:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    pmxvf64gerpp acc0, vsp4, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gerpp(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc, i32 0, i32 0)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf64gerpp(<512 x i1>, <256 x i1>, <16 x i8>, i32, i32)
+
+define void @test40(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test40:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    pmxvf64gerpn acc0, vsp4, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test40:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    pmxvf64gerpn acc0, vsp4, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gerpn(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc, i32 0, i32 0)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf64gerpn(<512 x i1>, <256 x i1>, <16 x i8>, i32, i32)
+
+define void @test41(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test41:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    pmxvf64gernp acc0, vsp4, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test41:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    pmxvf64gernp acc0, vsp4, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gernp(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc, i32 0, i32 0)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf64gernp(<512 x i1>, <256 x i1>, <16 x i8>, i32, i32)
+
+define void @test42(i8* %vqp, i8* %vpp, <16 x i8> %vc, i8* %resp) {
+; CHECK-LABEL: test42:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lxv vs1, 32(r3)
+; CHECK-NEXT:    lxv vs0, 48(r3)
+; CHECK-NEXT:    lxv vs3, 0(r3)
+; CHECK-NEXT:    lxv vs2, 16(r3)
+; CHECK-NEXT:    xxmtacc acc0
+; CHECK-NEXT:    lxv vs5, 0(r4)
+; CHECK-NEXT:    lxv vs4, 16(r4)
+; CHECK-NEXT:    pmxvf64gernn acc0, vsp4, v2, 0, 0
+; CHECK-NEXT:    xxmfacc acc0
+; CHECK-NEXT:    stxv vs0, 48(r7)
+; CHECK-NEXT:    stxv vs1, 32(r7)
+; CHECK-NEXT:    stxv vs2, 16(r7)
+; CHECK-NEXT:    stxv vs3, 0(r7)
+; CHECK-NEXT:    blr
+;
+; CHECK-BE-LABEL: test42:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxv vs1, 16(r3)
+; CHECK-BE-NEXT:    lxv vs0, 0(r3)
+; CHECK-BE-NEXT:    lxv vs3, 48(r3)
+; CHECK-BE-NEXT:    lxv vs2, 32(r3)
+; CHECK-BE-NEXT:    xxmtacc acc0
+; CHECK-BE-NEXT:    lxv vs5, 16(r4)
+; CHECK-BE-NEXT:    lxv vs4, 0(r4)
+; CHECK-BE-NEXT:    pmxvf64gernn acc0, vsp4, v2, 0, 0
+; CHECK-BE-NEXT:    xxmfacc acc0
+; CHECK-BE-NEXT:    stxv vs1, 16(r7)
+; CHECK-BE-NEXT:    stxv vs0, 0(r7)
+; CHECK-BE-NEXT:    stxv vs3, 48(r7)
+; CHECK-BE-NEXT:    stxv vs2, 32(r7)
+; CHECK-BE-NEXT:    blr
+entry:
+  %0 = bitcast i8* %vqp to <512 x i1>*
+  %1 = load <512 x i1>, <512 x i1>* %0, align 64
+  %2 = bitcast i8* %vpp to <256 x i1>*
+  %3 = load <256 x i1>, <256 x i1>* %2, align 32
+  %4 = tail call <512 x i1> @llvm.ppc.mma.pmxvf64gernn(<512 x i1> %1, <256 x i1> %3, <16 x i8> %vc, i32 0, i32 0)
+  %5 = bitcast i8* %resp to <512 x i1>*
+  store <512 x i1> %4, <512 x i1>* %5, align 64
+  ret void
+}
+
+
+declare <512 x i1> @llvm.ppc.mma.pmxvf64gernn(<512 x i1>, <256 x i1>, <16 x i8>, i32, i32)