tilelang.intrinsics.mfma_layout module#
- tilelang.intrinsics.mfma_layout.make_mfma_swizzle_layout(shared_buf, vecSize=8)#
- tilelang.intrinsics.mfma_layout.shared_16x16_to_ldmatrix_64x4_layout(ind)#
- tilelang.intrinsics.mfma_layout.shared_16x16_to_local_64x4_layout_A(i, j)#
- tilelang.intrinsics.mfma_layout.shared_16x16_to_local_64x4_layout_B(i, j)#
- tilelang.intrinsics.mfma_layout.shared_16x16_to_local_64x4_layout_C(i, j)#
- tilelang.intrinsics.mfma_layout.shared_16x32_to_local_64x8_layout_A(i, j)#
- tilelang.intrinsics.mfma_layout.shared_16x32_to_local_64x8_layout_B(i, j)#
- tilelang.intrinsics.mfma_layout.shared_16x4_to_local_64x1_layout_A(i, j)#
- tilelang.intrinsics.mfma_layout.shared_4x16_to_local_64x1_layout_B(i, j)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x1_to_16x4_layout_A(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x1_to_4x16_layout_B(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x4_to_16x16_layout_A(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x4_to_16x16_layout_B(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x4_to_16x16_layout_C_m_n(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x4_to_16x16_layout_C_n_m(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x8_to_16x32_layout_A(thread_id, local_id)#
- tilelang.intrinsics.mfma_layout.thread_id_shared_access_64x8_to_16x32_layout_B(thread_id, local_id)#