HomePhabricator

[CUDA] Implemented __nvvm_atom_*_gen_* builtins.

Description

[CUDA] Implemented __nvvm_atom_*_gen_* builtins.

Integer variants are implemented as atomicrmw or cmpxchg instructions.
Atomic add for floating point (__nvvm_atom_add_gen_f()) is implemented
as a call to an overloaded @llvm.nvvm.atomic.load.add.f32.* LVVM
intrinsic.

Differential Revision: http://reviews.llvm.org/D10666

Details

Committed
traJun 25 2015, 11:29 AM
Differential Revision
D10666: [CUDA] Implemented __nvvm_atom_*_gen_* builtins.
Branches
Unknown
Tags
Unknown