Script.atomic.shared_scatter_max

Script.atomic.shared_scatter_max(dst, *, dim, indices, values, sem='relaxed', scope='cta', output=None)

Scatter-max into a shared tile along dim.

See shared_scatter_add() for the parameter description.

Notes

• Thread group: Can be executed by any sized thread group.

• Hardware: Requires compute capability 7.0+ (sm_70).

• PTX: atom.{sem}.{scope}.shared.max.s32.

Parameters:

• dst (SharedTensor)

• dim (int)

• indices (RegisterTensor)

• values (RegisterTensor)

• sem (str)

• scope (str)

• output (RegisterTensor | None)

Return type:

RegisterTensor | None