tilelang.tileop.gemm
====================

.. py:module:: tilelang.tileop.gemm


Submodules
----------

.. toctree::
   :maxdepth: 1

   /autoapi/tilelang/tileop/gemm/gemm_base/index
   /autoapi/tilelang/tileop/gemm/gemm_mfma/index
   /autoapi/tilelang/tileop/gemm/gemm_mma/index
   /autoapi/tilelang/tileop/gemm/gemm_mma_sm70/index
   /autoapi/tilelang/tileop/gemm/gemm_tcgen05/index
   /autoapi/tilelang/tileop/gemm/gemm_wgmma/index


Classes
-------

.. autoapisummary::

   tilelang.tileop.gemm.GemmInst
   tilelang.tileop.gemm.GemmPy


Functions
---------

.. autoapisummary::

   tilelang.tileop.gemm.gemm_py_infer_layout
   tilelang.tileop.gemm.gemm_py_lower


Package Contents
----------------

.. py:function:: gemm_py_infer_layout(gemm_py, target, thread_bounds)

.. py:function:: gemm_py_lower(gemm_py, layout_map, target, thread_bounds, thread_var)

.. py:class:: GemmInst

   Bases: :py:obj:`enum.IntEnum`


   Enum where members are also (and must be) ints


   .. py:attribute:: MMA
      :value: 0



   .. py:attribute:: WGMMA
      :value: 1



   .. py:attribute:: TCGEN5MMA
      :value: 2



   .. py:attribute:: MFMA
      :value: 3



   .. py:method:: is_mma()


   .. py:method:: is_wgmma()


   .. py:method:: is_tcgen5mma()


   .. py:method:: is_mfma()


   .. py:method:: __repr__()

      Return repr(self).



.. py:class:: GemmPy

   Bases: :py:obj:`tvm.ir.base.Node`, :py:obj:`tvm.runtime.Scriptable`


   .. py:property:: A


   .. py:property:: B


   .. py:property:: C


   .. py:property:: APtr


   .. py:property:: BPtr


   .. py:property:: CPtr


   .. py:property:: M


   .. py:property:: N


   .. py:property:: K


   .. py:property:: trans_A


   .. py:property:: trans_B


   .. py:property:: stride_A


   .. py:property:: stride_B


   .. py:property:: offset_A


   .. py:property:: offset_B


   .. py:property:: clear_accum


   .. py:property:: k_pack


   .. py:property:: wg_wait


   .. py:method:: infer_layout(target, thread_nums)

      Infer the layout for the GEMM operation based on target architecture.



   .. py:method:: lower(layout_map, target, thread_nums, thread_var)

      Lower the GEMM operation to TIR statements based on target architecture.