The nvmath.device module is experimental and potentially subject to future changes.

TilePipeline#

class nvmath.device.TilePipeline(device_pipeline: DevicePipeline)[source]#

TilePipeline allows users to execute a pipelined matrix multiplication with partial tile results accumulated into an accumulator.

Refer to the cuBLASDx documentation for more details on how to use this class: using_pipelines.html

Attributes

storage_alignment#
storage_bytes#

Methods

execute(accumulator)[source]#