[X86] Add builtins for vector element insert and extract for different 128 and 256 bit vector types. Use them to implement the extract and insert intrinsics.
Previously we were just using extended vector operations in the header file.
This unfortunately allowed non-constant indices to be used with the intrinsics. This is incompatible with gcc, icc, and MSVC. It also introduces a different performance characteristic because non-constant index gets lowered to a vector store and an element sized load.
By adding the builtins we can check for the index to be a constant and ensure its in range of the vector element count.
User code still has the option to use extended vector operations themselves if they need non-constant indexing.
llvm-svn: 334057
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 6379f53..ad2e1cf 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -8778,8 +8778,29 @@
return Builder.CreateBitCast(BuildVector(Ops),
llvm::Type::getX86_MMXTy(getLLVMContext()));
case X86::BI__builtin_ia32_vec_ext_v2si:
- return Builder.CreateExtractElement(Ops[0],
- cast<ConstantInt>(Ops[1])->getZExtValue());
+ case X86::BI__builtin_ia32_vec_ext_v16qi:
+ case X86::BI__builtin_ia32_vec_ext_v8hi:
+ case X86::BI__builtin_ia32_vec_ext_v4si:
+ case X86::BI__builtin_ia32_vec_ext_v4sf:
+ case X86::BI__builtin_ia32_vec_ext_v2di:
+ case X86::BI__builtin_ia32_vec_ext_v32qi:
+ case X86::BI__builtin_ia32_vec_ext_v16hi:
+ case X86::BI__builtin_ia32_vec_ext_v8si:
+ case X86::BI__builtin_ia32_vec_ext_v4di:
+ // These builtins exist so we can ensure the index is an ICE and in range.
+ // Otherwise we could just do this in the header file.
+ return Builder.CreateExtractElement(Ops[0], Ops[1]);
+ case X86::BI__builtin_ia32_vec_set_v16qi:
+ case X86::BI__builtin_ia32_vec_set_v8hi:
+ case X86::BI__builtin_ia32_vec_set_v4si:
+ case X86::BI__builtin_ia32_vec_set_v2di:
+ case X86::BI__builtin_ia32_vec_set_v32qi:
+ case X86::BI__builtin_ia32_vec_set_v16hi:
+ case X86::BI__builtin_ia32_vec_set_v8si:
+ case X86::BI__builtin_ia32_vec_set_v4di:
+ // These builtins exist so we can ensure the index is an ICE and in range.
+ // Otherwise we could just do this in the header file.
+ return Builder.CreateInsertElement(Ops[0], Ops[1], Ops[2]);
case X86::BI_mm_setcsr:
case X86::BI__builtin_ia32_ldmxcsr: {
Address Tmp = CreateMemTemp(E->getArg(0)->getType());