[PATCH] D18592: [PowerPC] Back end improvements to vec_splat

[Nemanja Ivanovic via llvm-commits]

nemanjai created this revision.
nemanjai added reviewers: hfinkel, kbarton, wschmidt, amehsan, seurer.
nemanjai added a subscriber: llvm-commits.
nemanjai set the repository for this revision to rL LLVM.

We currently have no way to emit the xxspltw for a word splat and a call to vec_splat for vectors with 8-byte elements get translated to vperm with a mask vector that comes from memory.
This patch provides intrinsics to get at xxspltw and xxspltd (extended mnemonic). In a subsequent patch, altivec.h will be modified to use these intrinsics for the respective vec_splat definitions.

This provides a significant improvement in one benchmark that uses vec_splat.

Repository:
  rL LLVM

http://reviews.llvm.org/D18592

Files:
  include/llvm/IR/IntrinsicsPowerPC.td
  lib/Target/PowerPC/PPCInstrVSX.td

Index: lib/Target/PowerPC/PPCInstrVSX.td
===================================================================
--- lib/Target/PowerPC/PPCInstrVSX.td
+++ lib/Target/PowerPC/PPCInstrVSX.td
@@ -765,7 +765,9 @@
 
   def XXPERMDI : XX3Form_2<60, 10,
                        (outs vsrc:$XT), (ins vsrc:$XA, vsrc:$XB, u2imm:$DM),
-                       "xxpermdi $XT, $XA, $XB, $DM", IIC_VecPerm, []>;
+                       "xxpermdi $XT, $XA, $XB, $DM", IIC_VecPerm,
+                       [(set v2i64:$XT, (int_ppc_vsx_xxpermdi v2i64:$XA,
+                                          v2i64:$XB, imm:$DM))]>;
   def XXSEL : XX4Form<60, 3,
                       (outs vsrc:$XT), (ins vsrc:$XA, vsrc:$XB, vsrc:$XC),
                       "xxsel $XT, $XA, $XB, $XC", IIC_VecPerm, []>;
@@ -775,7 +777,9 @@
                        "xxsldwi $XT, $XA, $XB, $SHW", IIC_VecPerm, []>;
   def XXSPLTW : XX2Form_2<60, 164,
                        (outs vsrc:$XT), (ins vsrc:$XB, u2imm:$UIM),
-                       "xxspltw $XT, $XB, $UIM", IIC_VecPerm, []>;
+                       "xxspltw $XT, $XB, $UIM", IIC_VecPerm,
+                       [(set v4i32:$XT,
+                             (int_ppc_vsx_xxspltw v4i32:$XB, imm:$UIM))]>;
 } // hasSideEffects
 
 // SELECT_CC_* - Used to implement the SELECT_CC DAG operation.  Expanded after
Index: include/llvm/IR/IntrinsicsPowerPC.td
===================================================================
--- include/llvm/IR/IntrinsicsPowerPC.td
+++ include/llvm/IR/IntrinsicsPowerPC.td
@@ -746,6 +746,13 @@
 def int_ppc_vsx_xxleqv :
       PowerPC_VSX_Intrinsic<"xxleqv", [llvm_v4i32_ty],
                             [llvm_v4i32_ty, llvm_v4i32_ty], [IntrNoMem]>;
+def int_ppc_vsx_xxspltw :
+      PowerPC_VSX_Intrinsic<"xxspltw", [llvm_v4i32_ty],
+                            [llvm_v4i32_ty, llvm_i32_ty], [IntrNoMem]>;
+def int_ppc_vsx_xxpermdi :
+      PowerPC_VSX_Intrinsic<"xxpermdi", [llvm_v2i64_ty],
+                            [llvm_v2i64_ty, llvm_v2i64_ty, llvm_i32_ty],
+                            [IntrNoMem]>;
 }
 
 //===----------------------------------------------------------------------===//


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D18592.52035.patch
Type: text/x-patch
Size: 2147 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160330/a55a0621/attachment.bin>