[PATCH] D37985: [AMDGPU] add LDS f32 intrinsics
Daniil Fukalov via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon Sep 18 11:42:44 PDT 2017
dfukalov marked an inline comment as done.
dfukalov added inline comments.
================
Comment at: include/llvm/IR/IntrinsicsAMDGPU.td:303
+class AMDGPUAtomicF32IntrinNORET : Intrinsic<[],
+ [LLVMQualPointerType<llvm_float_ty, 3>, llvm_float_ty],
+ [IntrArgMemOnly, NoCapture<0>, IntrNoReturn]
----------------
arsenm wrote:
> Should this have an operand added for the ordering?
No, these intrinsics are created by request to be able to generate ds_{add|min|max}[_rtn]_f32 in case of OpenCL local memory atomics only. They work only for pointers to floats located in addrspace 3
================
Comment at: include/llvm/IR/IntrinsicsAMDGPU.td:304
+ [LLVMQualPointerType<llvm_float_ty, 3>, llvm_float_ty],
+ [IntrArgMemOnly, NoCapture<0>, IntrNoReturn]
+>;
----------------
arsenm wrote:
> This is certainly not IntrNoReturn
You're right, I interpreted it as intrinsic that has no return value, diff updated
https://reviews.llvm.org/D37985
More information about the llvm-commits
mailing list