triton.experimental.gluon.language.nvidia.blackwell.tensor_memory_descriptor
- class triton.experimental.gluon.language.nvidia.blackwell.tensor_memory_descriptor(self, handle, element_ty, shape, layout, alloc_shape)
Represents a tensor memory descriptor handle for Tensor Core Gen5 operations.
- __init__(self, handle, element_ty, shape, layout, alloc_shape)
Methods
__init__(self, handle, element_ty, shape, ...)get_reg_layout(self[, num_warps, ...])Return the register layout used to access this tensor memory descriptor.
index(self, index[, _semantic])Create a subview of tensor memory by indexing the first dimension.
load(self[, layout, _semantic, _generator])Load a tensor from tensor memory.
load_max(self[, layout, abs, propagate_nan, ...])Load a tensor from tensor memory with MAX reduction along the N-dimension.
load_min(self[, layout, abs, propagate_nan, ...])Load a tensor from tensor memory with MIN reduction along the N-dimension.
slice(self, start, length[, _semantic])Create a slice of the tensor memory descriptor along the last dimension.
store(self, value[, pred, _semantic])Store a tensor into tensor memory.
Attributes
dtypelayoutrankshapetype