IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /max/get-started.md). For the complete documentation index, see llms.txt.
Skip to main content
For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo function

blackwell_tma_umma_warp_specialized_kernel

def blackwell_tma_umma_warp_specialized_kernel[a_type: DType, b_type: DType, c_type: DType, expert_m: Int, a_tile_rank: Int, a_tile_shape: IndexList[a_tile_rank], a_desc_shape: IndexList[a_tile_rank], b_tile_rank: Int, b_tile_shape: IndexList[b_tile_rank], b_desc_shape: IndexList[b_tile_rank], c_tile_rank: Int, c_tile_shape_param: IndexList[c_tile_rank], c_desc_shape: IndexList[c_tile_rank], block_tile_shape: IndexList[Int(3)], mma_shape: IndexList[Int(3)], cluster_shape: StaticTuple[Int32, Int(3)], num_pipeline_stages: Int, num_accum_pipeline_stages: Int, num_output_stages: Int = Int(2), output_tile_shape: IndexList[Int(2)] = Index[Int, Int](Int(128), Int(32)), transpose_b: Bool = True, a_swizzle: TensorMapSwizzle = TensorMapSwizzle.SWIZZLE_128B, b_swizzle: TensorMapSwizzle = TensorMapSwizzle.SWIZZLE_128B, c_swizzle: TensorMapSwizzle = TensorMapSwizzle.SWIZZLE_128B, cta_group: Int = Int(2), elementwise_lambda_fn: Optional[def[dtype: DType, width: SIMDSize, *, alignment: Int = Int(1)](IndexList[Int(2)], SIMD[dtype, width]) capturing -> None] = None, transpose_c: Bool = False, use_tma: Bool = True, K_actual: Int = Int(0), a_gmem_layout: Layout = Layout.row_major(Int(1), Int(1)), b_gmem_layout: Layout = Layout.row_major(Int(1), Int(1))](expert_usage_stats: UnsafePointer[UInt32, ImmutAnyOrigin], a_tma_op: TMATensorTile[a_type, a_tile_rank, a_tile_shape, a_desc_shape], expert_ids: UnsafePointer[Int32, ImmutAnyOrigin], b_tma_op: TMATensorTile[b_type, b_tile_rank, b_tile_shape, b_desc_shape], b_offsets: UnsafePointer[UInt32, ImmutAnyOrigin], c_tma_op: TMATensorTile[c_type, c_tile_rank, c_tile_shape_param, c_desc_shape], c_ptr: UnsafePointer[Scalar[c_type], MutAnyOrigin], mnk: StaticTuple[UInt32, Int(3)], a_gmem: LayoutTensor[a_type, a_gmem_layout, ImmutAnyOrigin], b_gmem: LayoutTensor[b_type, b_gmem_layout, ImmutAnyOrigin])