[LLVMdev] processor Intrisic handling question (original) (raw)
Stavropoulos Nikos n.stavropoulos at think-silicon.com
Fri Oct 24 03:11:10 PDT 2014
- Previous message: [LLVMdev] AArch64 bot
- Next message: [LLVMdev] Adding masked vector load and store intrinsics
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
hi all, i have 2 intrinsics that i want to insert them into my C code.
def int_nema_lkp3: GCCBuiltin<"__builtin_nema_lkp3">,Intrinsic<[llvm_v4f32_ty], [llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty]>;
def int_nema_updt: GCCBuiltin<"__builtin_nema_updt">,Intrinsic<[], [llvm_v4f32_ty]>;
and the C sample code is
main {
float4 a,b,c,d;
d=__builtin_nema_lkp3(a,b,c); /// if have hit jump after __builtin_nema_updt
d= a*b+c;
__builtin_nema_updt(d);
return 0; }
so i want the first intrinsic instruction to conditionally branch after the second intrinsic instruction.
i have tried to do the same with with if - else code but this needs an extra variable that i don't want to exist.
Should i do something like iterating the nodes to find where to jump??
Thanks in advance
-- View this message in context: http://llvm.1065342.n5.nabble.com/LLVMdev-processor-Intrisic-handling-question-tp73755.html Sent from the LLVM - Dev mailing list archive at Nabble.com.
- Previous message: [LLVMdev] AArch64 bot
- Next message: [LLVMdev] Adding masked vector load and store intrinsics
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]