triton.experimental.gluon.language.nvidia.hopper.tma.async_store

triton.experimental.gluon.language.nvidia.hopper.tma.async_store(tensor_desc, coord, src, cache_policy=None, _semantic=None)

Store data from shared memory to global memory using TMA.

Parameters:

  • tensor_desc (tensor_descriptor) – Tensor descriptor (tiled).

  • coord (Sequence[int | ttgl.constexpr | ttgl.tensor]) – Coordinates in the destination tensor.

  • src (ttgl.shared_memory_descriptor) – Source memory descriptor.

  • cache_policy – A compile-time CachePolicy containing only an L2 eviction policy. None leaves the policy unspecified. Cache policies are performance hints and do not affect synchronization.