################################################################################################# # Copyright (c) 2023 - 2025 Hygon Information Technology Co., Ltd. All rights reserved. # SPDX-License-Identifier: BSD-3-Clause # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions are met: # # 1. Redistributions of source code must retain the above copyright notice, this # list of conditions and the following disclaimer. # # 2. Redistributions in binary form must reproduce the above copyright notice, # this list of conditions and the following disclaimer in the documentation # and/or other materials provided with the distribution. # # 3. Neither the name of the copyright holder nor the names of its # contributors may be used to endorse or promote products derived from # this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. # ################################################################################################# from itertools import product from copy import deepcopy try: import builtins if hasattr(builtins, "HYTLASS_IGNORE_PACKAGE") and HYTLASS_IGNORE_PACKAGE == True: raise ImportError("Disabling attempt to import hytlass_library") from hytlass_library.library import * except ImportError: from library import * class TileConfig: device_config_params = { "Gfx906": { "min_cc": 906, "max_cc": 1024, "smem_size": 65536, "stages": [2], "cluster_shapes": [[1,1,1]], "warp_count_mapping":{ 64: [[1, 1, 1]], 128: [[2, 1, 1], [1, 2, 1]], 256: [[2, 2, 1], [1, 4, 1], [4, 1, 1]] }, "schedules": [ [KernelScheduleType.ScheduleAuto, EpilogueScheduleType.ScheduleAuto] ] }, "Gfx928": { "min_cc": 928, "max_cc": 1024, "smem_size": 65536, "stages": [2], "cluster_shapes": [[1,1,1]], "warp_count_mapping":{ 64: [[1, 1, 1]], 128: [[2, 1, 1], [1, 2, 1]], 256: [[2, 2, 1], [1, 4, 1], [4, 1, 1]], 512: [[2, 4, 1], [4, 2, 1], [1, 8, 1], [8, 1, 1]] }, "schedules": [ [KernelScheduleType.ScheduleAuto, EpilogueScheduleType.ScheduleAuto], [KernelScheduleType.ScheduleAuto, EpilogueScheduleType.NoSmemWarpSpecialized] ], "stream_k_schedules": [ [KernelScheduleType.StreamK, EpilogueScheduleType.ScheduleAuto] ] } } # 各设备不同数据类型math_instructions data_type_math_instructions_params = { "Gfx906": { "8b": [ MathInstruction( [1, 1, 4], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.Simt, MathOperation.multiply_add), ], "16b": [ MathInstruction( [1, 1, 2], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.Simt, MathOperation.multiply_add), ], "32b": [ MathInstruction( [1, 1, 1], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.Simt, MathOperation.multiply_add), ] }, "Gfx928": { "8b": [ MathInstruction( [32, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [32, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ] } } # aligment 配置参数 data_type_aligment_params = { "8b": [16], "16b": [8], "32b": [4] } # 这里只提供Layout类型，具体的对齐参数在后续设置 layout_mappings = { "nt": [[LayoutType.ColumnMajor, 1], [LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1]], "nn": [[LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1]], "tt": [[LayoutType.RowMajor, 1], [LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1]], "tn": [[LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1]], } instruction_mappings = { "nt": { "16b": [ MathInstruction( [32, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [32, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "nn": { "16b": [ MathInstruction( [32, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add) ], "8b": [ MathInstruction( [32, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "tt": { "16b": [ MathInstruction( [32, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [32, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "tn": { "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 64], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 64], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 64], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 64], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 64], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 32, 64], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 32, 64], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [32, 16, 64], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], } } def __init__(self, device_name, data_type, layout=""): # 根据设备名称获取设备相关配置参数 device_config = self.device_config_params.get(device_name) if device_config is None: raise ValueError(f"Unsupported device: {device_name}") self.min_cc = device_config["min_cc"] self.max_cc = device_config["max_cc"] self.smem_size = device_config["smem_size"] self.stages = device_config["stages"] self.cluster_shapes = device_config["cluster_shapes"] self.warp_count_mapping = device_config["warp_count_mapping"] self.schedules = device_config["schedules"] self.stream_k_schedules = device_config.get("stream_k_schedules", []) # 根据数据类型获取指令配置参数 if layout != "" : if layout not in ["nt", "nn", "tn", "tt"]: raise ValueError(f"Unsupported layout type: {layout}") math_instructions = self.instruction_mappings.get(layout).get(data_type) else : math_instructions = self.data_type_math_instructions_params.get(device_name,{}).get(data_type) if math_instructions is None: raise ValueError(f"Unsupported data type: {data_type}") self.math_instructions = math_instructions # 根据数据类型获取对齐配置参数 data_type_aligment = self.data_type_aligment_params.get(data_type) if data_type_aligment is None: raise ValueError(f"Unsupported data type: {data_type}") if layout != "" : self.current_layouts = [] for alignment in data_type_aligment: raw_layout = deepcopy(self.layout_mappings.get(layout)) raw_layout[0][1] = alignment raw_layout[1][1] = alignment self.current_layouts.append(raw_layout) else : self.current_layouts = self.layout_aligment_config(data_type_aligment) def layout_aligment_config(self, data_type_aligment): layouts = [ [[LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1]], [[LayoutType.ColumnMajor, 1], [LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1]], [[LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1], [LayoutType.ColumnMajor, 1]], [[LayoutType.RowMajor, 1], [LayoutType.RowMajor, 1], [LayoutType.ColumnMajor, 1]], ] if data_type_aligment : current_layouts = [] for aligment in data_type_aligment: for layout in layouts: new_layout = [list(layout[0]), list(layout[1]), list(layout[2])] new_layout[0][1] = aligment new_layout[1][1] = aligment current_layouts.append(new_layout) return current_layouts else: return layouts class TileConfig_2x: # Gfx906、Gfx928设备相关配置参数：min_cc、max_cc、smem_size、stages、cluster_shapes device_config_params = { "Gfx928": { "min_cc": 928, "max_cc": 1024, "smem_size": 65536, # 主要对单流水支持 "stages": [1], "warp_count_mapping":{ 64: [[1, 1, 1]], 128: [[2, 1, 1], [1, 2, 1], [1, 1, 2]], 256: [[2, 2, 1], [1, 4, 1], [4, 1, 1], [2, 1, 2], [1, 2, 2], [1, 1, 4]], 512: [[2, 4, 1], [4, 2, 1], [1, 8, 1], [8, 1, 1], [4, 1, 2], [2, 2, 2], [1, 4, 2]] }, } } # 不同数据类型 math_instructions data_type_math_instructions_params = { "Gfx928": { "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ] } } # aligment 配置参数 data_type_aligment_params = { "8b" : [[16, 16, 8]], "16b": [[8, 8, 8]], "32b": [[4, 4, 4]] } # 这里只提供Layout类型，具体的对齐参数在后续设置 layout_mappings = { "nt" : (LayoutType.ColumnMajor, LayoutType.RowMajor, LayoutType.ColumnMajor), "tt" : (LayoutType.RowMajor, LayoutType.RowMajor, LayoutType.ColumnMajor), "tn" : (LayoutType.RowMajor, LayoutType.ColumnMajor, LayoutType.ColumnMajor), "nn" : (LayoutType.ColumnMajor, LayoutType.ColumnMajor, LayoutType.ColumnMajor), } instruction_mappings = { "nt": { "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "tt": { "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "nn": { "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, "tn": { "32b": [ MathInstruction( [16, 16, 8], DataType.f32, DataType.f32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 8], DataType.tf32, DataType.tf32, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "16b": [ MathInstruction( [16, 16, 16], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 16], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.f16, DataType.f16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.bf16, DataType.bf16, DataType.f32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], "8b": [ MathInstruction( [16, 16, 32], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 32], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 64], DataType.s8, DataType.s8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), MathInstruction( [16, 16, 64], DataType.u8, DataType.u8, DataType.s32, OpcodeClass.TensorOp, MathOperation.multiply_add), ], }, } def __init__(self, device_name, data_type, layout=""): # 根据设备名称获取设备相关配置参数 device_config = self.device_config_params.get(device_name) if device_config is None: raise ValueError(f"Unsupported device: {device_name}") self.min_cc = device_config["min_cc"] self.max_cc = device_config["max_cc"] self.smem_size = device_config["smem_size"] self.stages = device_config["stages"] self.warp_count_mapping = device_config["warp_count_mapping"] # 根据数据类型获取指令配置参数 if layout != "" : if layout not in ["nt", "nn", "tn", "tt"]: raise ValueError(f"Unsupported layout type: {layout}") math_instructions = self.instruction_mappings.get(layout).get(data_type) else : math_instructions = self.data_type_math_instructions_params.get(device_name,{}).get(data_type) if math_instructions is None: raise ValueError(f"Unsupported data type: {data_type}") self.math_instructions = math_instructions self.data_type_aligment = self.data_type_aligment_params.get(data_type) if self.data_type_aligment is None: raise ValueError(f"Unsupported data type: {data_type}") self.layouts = [] if layout != "": self.layouts.append(self.layout_mappings.get(layout)) else: for _layout in range(["nt", "tn", "nn", "tt"]): self.layouts.append(self.layout_mappings.get(_layout))