[autoparallel] support addmm in tracer and solver (#1961)

* [fx] patch addmm

* [autoparallel] support addmm in tracer and solver

colossalai/auto_parallel/tensor_shard/node_handler/__init__.py

@@ -1,3 +1,4 @@
+from .addmm_handler import ADDMMFunctionHandler
 from .batch_norm_handler import BatchNormModuleHandler
 from .binary_elementwise_handler import BinaryElementwiseHandler
 from .bmm_handler import AddBMMFunctionHandler, BMMFunctionHandler
@@ -18,5 +19,5 @@ __all__ = [
     'LinearFunctionHandler', 'LinearModuleHandler', 'BMMFunctionHandler', 'AddBMMFunctionHandler',
     'LayerNormModuleHandler', 'BatchNormModuleHandler', 'ConvModuleHandler', 'ConvFunctionHandler',
     'UnaryElementwiseHandler', 'ReshapeHandler', 'PlacehodlerHandler', 'OuputHandler', 'WhereHandler',
-    'NormPoolingHandler', 'BinaryElementwiseHandler', 'MatMulHandler', 'operator_registry'
+    'NormPoolingHandler', 'BinaryElementwiseHandler', 'MatMulHandler', 'operator_registry', 'ADDMMFunctionHandler'
 ]

colossalai/auto_parallel/tensor_shard/node_handler/addmm_handler.py

@@ -0,0 +1,91 @@
+from typing import Dict, List, Union
+
+import torch
+
+from colossalai.tensor.shape_consistency import CollectiveCommPattern, CommSpec, ShapeConsistencyManager
+
+from ..sharding_strategy import CommAction, CommType, OperationData, OperationDataType, ShardingStrategy
+from ..utils import comm_actions_for_oprands, recover_sharding_spec_for_broadcast_shape
+from .node_handler import NodeHandler
+from .registry import operator_registry
+from .strategy import LinearProjectionStrategyGenerator, StrategyGenerator
+
+__all__ = ['ADDMMFunctionHandler']
+
+
+@operator_registry.register(torch.addmm)
+@operator_registry.register(torch.Tensor.addmm)
+class ADDMMFunctionHandler(NodeHandler):
+    """
+    This is a NodeHandler class which deals with the batched matrix multiplication operation in PyTorch.
+    Such operations including `torch.bmm` and `torch.Tensor.bmm` require the tensor to be 3D, thus, there is
+    no logical-physical shape conversion in this handler.
+    """
+
+    def _infer_op_data_type(self, tensor: torch.Tensor) -> OperationDataType:
+        if isinstance(tensor, torch.nn.parameter.Parameter):
+            data_type = OperationDataType.PARAM
+        else:
+            data_type = OperationDataType.ARG
+        return data_type
+
+    def get_operation_data_mapping(self) -> Dict[str, OperationData]:
+
+        # input operand
+        input_data = self.node.args[1]._meta_data
+        physical_input_operand = OperationData(name=str(self.node.args[1]),
+                                               type=self._infer_op_data_type(input_data),
+                                               data=input_data)
+
+        # other operand
+        other_data = self.node.args[2]._meta_data
+        physical_other_operand = OperationData(name=str(self.node.args[2]),
+                                               type=self._infer_op_data_type(other_data),
+                                               data=other_data)
+        # bias physical shape
+        bias_logical_shape = self.node._meta_data.shape
+        bias_data = self.node.args[0]._meta_data
+        physical_bias_operand = OperationData(name=str(self.node.args[0]),
+                                              type=self._infer_op_data_type(bias_data),
+                                              data=bias_data,
+                                              logical_shape=bias_logical_shape)
+
+        # output
+        physical_output = OperationData(name=str(self.node), type=OperationDataType.OUTPUT, data=self.node._meta_data)
+
+        mapping = {
+            "input": physical_input_operand,
+            "other": physical_other_operand,
+            "output": physical_output,
+            'bias': physical_bias_operand
+        }
+
+        return mapping
+
+    def get_strategy_generator(self) -> List[StrategyGenerator]:
+        op_data_mapping = self.get_operation_data_mapping()
+        generators = []
+        generators.append(
+            LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh, linear_projection_type='addmm'))
+        return generators
+
+    def post_process(self, strategy: ShardingStrategy) -> Union[ShardingStrategy, List[ShardingStrategy]]:
+        # convert bias from its logical sharding spec to its physical sharding spec
+        op_data_mapping = self.get_operation_data_mapping()
+
+        bias_op_data = op_data_mapping['bias']
+        bias_physical_shape = bias_op_data.data.shape
+        bias_logical_shape = bias_op_data.logical_shape
+        bias_sharding_spec = strategy.get_sharding_spec_by_name(bias_op_data.name)
+        bias_sharding_spec, removed_dims = recover_sharding_spec_for_broadcast_shape(
+            bias_sharding_spec, bias_logical_shape, bias_physical_shape)
+        strategy.sharding_specs[bias_op_data] = bias_sharding_spec
+
+        if len(removed_dims) > 0:
+            comm_action = comm_actions_for_oprands(node=self.node,
+                                                   removed_dims=removed_dims,
+                                                   op_data=bias_op_data,
+                                                   sharding_spec=bias_sharding_spec)
+            strategy.communication_actions[bias_op_data] = comm_action
+
+        return strategy

colossalai/auto_parallel/tensor_shard/node_handler/linear_handler.py

@@ -140,7 +140,8 @@ class LinearModuleHandler(ModuleHandler):
     def get_strategy_generator(self) -> List[StrategyGenerator]:
         op_data_mapping = self.get_operation_data_mapping()
         generators = []
-        generators.append(LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh))
+        generators.append(
+            LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh, linear_projection_type='linear'))
         return generators
 
     def get_operation_data_mapping(self) -> Dict[str, OperationData]:
@@ -199,7 +200,8 @@ class LinearFunctionHandler(NodeHandler):
     def get_strategy_generator(self) -> List[StrategyGenerator]:
         op_data_mapping = self.get_operation_data_mapping()
         generators = []
-        generators.append(LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh))
+        generators.append(
+            LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh, linear_projection_type='linear'))
         return generators
 
     def get_operation_data_mapping(self) -> Dict[str, OperationData]:

colossalai/auto_parallel/tensor_shard/node_handler/matmul_handler.py

@@ -363,7 +363,8 @@ class MatMulHandler(NodeHandler):
         elif self.matmul_type == MatMulType.MV:
             generators.append(MatVecStrategyGenerator(op_data_mapping, self.device_mesh))
         elif self.matmul_type == MatMulType.MM:
-            generators.append(LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh))
+            generators.append(
+                LinearProjectionStrategyGenerator(op_data_mapping, self.device_mesh, linear_projection_type='linear'))
         return generators
 
     def get_operation_data_mapping(self) -> Dict[str, OperationData]:

colossalai/auto_parallel/tensor_shard/node_handler/strategy/matmul_strategy_generator.py

@@ -209,6 +209,10 @@ class MatVecStrategyGenerator(MatMulStrategyGenerator):
 
 class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
 
+    def __init__(self, operation_data_mapping, device_mesh, linear_projection_type='linear'):
+        super().__init__(operation_data_mapping, device_mesh)
+        self.linear_projection_type = linear_projection_type
+
     def update_compute_cost(self, strategy: ShardingStrategy) -> ShardingStrategy:
         # C = AB
         # C: [M, N], A: [M, P], B: [P, N]
@@ -272,14 +276,21 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
             "other": {
                 -1: [mesh_dim_1]
             },
-            "bias": {
-                -1: [mesh_dim_1]
-            },
             "output": {
                 0: [mesh_dim_0],
                 -1: [mesh_dim_1]
             },
         }
+
+        # linear bias only has one dimension, but addmm bias has same dimensions
+        # as the output logically.
+        if self.linear_projection_type == 'linear':
+            dim_partition_dict_mapping['bias'] = {-1: [mesh_dim_1]}
+        elif self.linear_projection_type == 'addmm':
+            dim_partition_dict_mapping['bias'] = {0: [mesh_dim_0], -1: [mesh_dim_1]}
+        else:
+            raise ('Unsupported linear projection type')
+
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # set communication action
@@ -293,13 +304,13 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
 
         if self.is_param('other'):
             other_comm_action = self.get_communication_action(
-                sharding_spec_mapping["output"],
+                sharding_spec_mapping["other"],
                 communication_pattern=CollectiveCommPattern.IDENTITY_FWD_ALLREDUCE_BWD,
                 logical_process_axis=mesh_dim_0,
                 comm_type=CommType.HOOK)
         else:
             other_comm_action = self.get_communication_action(
-                sharding_spec_mapping["output"],
+                sharding_spec_mapping["other"],
                 communication_pattern=CollectiveCommPattern.IDENTITY_FWD_ALLREDUCE_BWD,
                 logical_process_axis=mesh_dim_0,
                 comm_type=CommType.BEFORE,
@@ -308,7 +319,9 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
         communication_action_mapping['input'] = input_comm_action
         communication_action_mapping['other'] = other_comm_action
 
-        if self.has_bias:
+        # we only add allreduce comm action for linear bias, because
+        # allreduce comm action for addmm bias will be considered in post processing
+        if self.has_bias and self.linear_projection_type == 'linear':
             if self.is_param('bias'):
                 bias_comm_action = self.get_communication_action(
                     sharding_spec_mapping["bias"],
@@ -347,6 +360,16 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 0: [mesh_dim_0]
             },
         }
+
+        # linear bias only has one dimension, but addmm bias has same dimensions
+        # as the output logically.
+        if self.linear_projection_type == 'linear':
+            dim_partition_dict_mapping['bias'] = {}
+        elif self.linear_projection_type == 'addmm':
+            dim_partition_dict_mapping['bias'] = {0: [mesh_dim_0]}
+        else:
+            raise ('Unsupported linear projection type')
+
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication action mapping
@@ -360,13 +383,13 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
 
         if self.is_param('other'):
             other_comm_action = self.get_communication_action(
-                sharding_spec_mapping["output"],
+                sharding_spec_mapping["other"],
                 communication_pattern=CollectiveCommPattern.IDENTITY_FWD_ALLREDUCE_BWD,
                 logical_process_axis=mesh_dim_0,
                 comm_type=CommType.HOOK)
         else:
             other_comm_action = self.get_communication_action(
-                sharding_spec_mapping["output"],
+                sharding_spec_mapping["other"],
                 communication_pattern=CollectiveCommPattern.IDENTITY_FWD_ALLREDUCE_BWD,
                 logical_process_axis=mesh_dim_0,
                 comm_type=CommType.BEFORE,
@@ -375,7 +398,9 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
         communication_action_mapping['other'] = other_comm_action
         communication_action_mapping['output'] = output_comm_action
 
-        if self.has_bias:
+        # we only add allreduce comm action for linear bias, because
+        # allreduce comm action for addmm bias will be considered in post processing
+        if self.has_bias and self.linear_projection_type == 'linear':
             if self.is_param('bias'):
                 bias_comm_action = self.get_communication_action(
                     sharding_spec_mapping["bias"],
@@ -415,6 +440,10 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 -1: [mesh_dim_1]
             },
         }
+
+        # We don't have to do anything special for bias here, because
+        # the bias is already the same sharding spec as the output.
+
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication actions
@@ -451,7 +480,8 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
             "bias": {},
             "output": {},
         }
-
+        # We don't have to do anything special for bias here, because
+        # the bias is already the same sharding spec as the output.
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication action
@@ -484,7 +514,8 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 -1: [mesh_dim]
             },
         }
-
+        # We don't have to do anything special for bias here, because
+        # the bias is already the same sharding spec as the output.
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication actions
@@ -515,6 +546,16 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 0: [mesh_dim_0, mesh_dim_1]
             },
         }
+
+        # linear bias only has one dimension, but addmm bias has same dimensions
+        # as the output logically.
+        if self.linear_projection_type == 'linear':
+            dim_partition_dict_mapping['bias'] = {}
+        elif self.linear_projection_type == 'addmm':
+            dim_partition_dict_mapping['bias'] = {0: [mesh_dim_0, mesh_dim_1]}
+        else:
+            raise ('Unsupported linear projection type')
+
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication action
@@ -534,7 +575,9 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 arg_index=1)
         communication_action_mapping['other'] = other_comm_action
 
-        if self.has_bias:
+        # we only add allreduce comm action for linear bias, because
+        # allreduce comm action for addmm bias will be considered in post processing
+        if self.has_bias and self.linear_projection_type == 'linear':
             if self.is_param('bias'):
                 bias_comm_action = self.get_communication_action(
                     sharding_spec=sharding_spec_mapping['bias'],
@@ -568,6 +611,9 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
             "bias": {},
             "output": {},
         }
+
+        # We don't have to do anything special for bias here, because
+        # the bias is already the same sharding spec as the output.
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication action
@@ -600,6 +646,9 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
                 -1: [mesh_dim_0, mesh_dim_1]
             },
         }
+
+        # We don't have to do anything special for bias here, because
+        # the bias is already the same sharding spec as the output.
         sharding_spec_mapping = self.to_sharding_spec_mapping(dim_partition_dict_mapping)
 
         # get communication action
@@ -626,10 +675,7 @@ class LinearProjectionStrategyGenerator(MatMulStrategyGenerator):
         assert input_data.data.dim() > 0 and other_data.data.dim() == 2
         assert other_data.logical_shape[0] == input_data.logical_shape[-1]
 
-        # check if bias has the same a valid dim
-        has_bias = "bias" in self.op_data
-
-        if has_bias:
+        if self.has_bias:
             bias_data = self.op_data['bias']
             assert bias_data.logical_shape[-1] == other_data.logical_shape[-1]
 

colossalai/fx/tracer/meta_patch/patched_function/arithmetic.py

@@ -72,11 +72,21 @@ def torch_linear(input, mat2, bias=None, *, out=None):
 def torch_addbmm(input, mat1, mat2, *, beta=1, alpha=1, out=None):
     if out is not None:
         raise ValueError("Don't support in-place abs for MetaTensor analysis")
-    batch_size, n, m = mat1.shape
+    _, n, _ = mat1.shape
     _, _, p = mat2.shape
     return torch.empty(n, p, device="meta")
 
 
+@meta_patched_function.register(torch.addmm)
+@meta_patched_function.register(torch.Tensor.addmm)
+def torch_addmm(input, mat1, mat2, *, beta=1, alpha=1, out=None):
+    if out is not None:
+        raise ValueError("Don't support in-place abs for MetaTensor analysis")
+    n, _ = mat1.shape
+    _, p = mat2.shape
+    return torch.empty(n, p, device="meta")
+
+
 @meta_patched_function.register(torch.var_mean)
 def torch_var_mean(input, dim, unbiased=True, keepdim=False, *, out=None):
     assert out is None, 'saving to out is not supported yet'

tests/test_auto_parallel/test_tensor_shard/test_node_handler/test_addmm_handler.py

@@ -0,0 +1,156 @@
+from faulthandler import disable
+from functools import partial
+from xml.dom import WrongDocumentErr
+
+import pytest
+import torch
+import torch.multiprocessing as mp
+import torch.nn as nn
+from typing_extensions import Self
+
+from colossalai.auto_parallel.tensor_shard.node_handler import ADDMMFunctionHandler
+from colossalai.auto_parallel.tensor_shard.sharding_strategy import (
+    OperationData,
+    OperationDataType,
+    ShardingStrategy,
+    StrategiesVector,
+)
+from colossalai.device.device_mesh import DeviceMesh
+from colossalai.fx import ColoGraphModule, ColoTracer
+from colossalai.initialize import launch
+from colossalai.logging import disable_existing_loggers
+from colossalai.testing import assert_close, parameterize, rerun_if_address_is_in_use
+from colossalai.testing.pytest_wrapper import run_on_environment_flag
+from colossalai.utils import free_port
+from tests.test_auto_parallel.test_tensor_shard.test_node_handler.utils import numerical_test_for_node_strategy
+
+
+class AddmmModel(nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    def forward(self, input, m1, m2):
+        x = torch.addmm(input, m1, m2)
+        return x
+
+
+def check_linear_function_handler(rank, input_shape, world_size, port):
+    disable_existing_loggers()
+    launch(config={}, rank=rank, world_size=world_size, host='localhost', port=port, backend='nccl')
+    model = AddmmModel().cuda()
+    physical_mesh_id = torch.arange(0, 4)
+    mesh_shape = (2, 2)
+    device_mesh = DeviceMesh(physical_mesh_id, mesh_shape, init_process_group=True)
+
+    input = torch.rand(input_shape).cuda()
+    m1 = torch.rand(4, 8).cuda()
+    m2 = torch.rand(8, 16).cuda()
+    # the index of addmm node in computation graph
+    node_index = 3
+    # strategy number of linear node
+    strategy_number = 10
+    # construct input args
+    input_args = [input, m1, m2]
+    # construct meta arg names
+    meta_arg_names = ['input', 'm1', 'm2']
+    numerical_test_for_node_strategy(model=model,
+                                     device_mesh=device_mesh,
+                                     node_index=node_index,
+                                     strategy_number=strategy_number,
+                                     input_args=input_args,
+                                     meta_arg_names=meta_arg_names)
+
+    tracer = ColoTracer()
+    graph = tracer.trace(model,
+                         meta_args={
+                             "input": torch.rand(input_shape).to('meta'),
+                             'm1': torch.rand(4, 8).to('meta'),
+                             'm2': torch.rand(8, 16).to('meta'),
+                         })
+    gm = ColoGraphModule(model, graph)
+    # [input_1, m1, m2, addmm, output]
+    node_list = list(graph.nodes)
+    addmm_node = node_list[3]
+    strategies_vector = StrategiesVector(addmm_node)
+
+    # build handler
+    handler = ADDMMFunctionHandler(node=addmm_node, device_mesh=device_mesh, strategies_vector=strategies_vector)
+
+    handler.register_strategy(compute_resharding_cost=False)
+    strategy_name_list = [val.name for val in strategies_vector]
+
+    # check operation data mapping
+    mapping = handler.get_operation_data_mapping()
+
+    assert mapping['input'].name == "m1"
+    assert mapping['input'].data.shape == torch.Size([4, 8])
+    assert mapping['input'].type == OperationDataType.ARG
+    assert mapping['input'].logical_shape == torch.Size([4, 8])
+
+    assert mapping['other'].name == "m2"
+    assert mapping['other'].data.shape == torch.Size([8, 16])
+    assert mapping['other'].type == OperationDataType.ARG
+    assert mapping['other'].logical_shape == torch.Size([8, 16])
+
+    assert mapping['bias'].name == "input_1"
+    assert mapping['bias'].data.shape == torch.Size(input_shape)
+    assert mapping['bias'].type == OperationDataType.ARG
+    assert mapping['bias'].logical_shape == torch.Size([4, 16])
+
+    assert mapping['output'].name == "addmm"
+    assert mapping['output'].data.shape == torch.Size([4, 16])
+    assert mapping['output'].type == OperationDataType.OUTPUT
+
+    # one strategy will be converted to different physical sharding spec
+    assert len(strategy_name_list) > 8
+
+    # SS = SR x RS
+    assert 'S0S1 = S0R x RS1' in strategy_name_list
+    assert 'S1S0 = S1R x RS0' in strategy_name_list
+
+    # SR = SS x SR
+    assert 'S0R = S0S1 x S1R' in strategy_name_list
+    assert 'S1R = S1S0 x S0R' in strategy_name_list
+
+    # RS = RS x SS
+    assert 'RS0 = RS1 x S1S0' in strategy_name_list
+    assert 'RS1 = RS0 x S0S1' in strategy_name_list
+
+    # RR = RS x SR
+    assert 'RR = RS0 x S0R' in strategy_name_list
+    assert 'RR = RS1 x S1R' in strategy_name_list
+
+    # RS= RR x RS
+    assert 'RS0 = RR x RS0' in strategy_name_list
+    assert 'RS1 = RR x RS1' in strategy_name_list
+
+    for strategy in strategies_vector:
+        strategy: ShardingStrategy
+        input_sharding_spec = strategy.get_sharding_spec_by_name('m1')
+        weight_sharding_spec = strategy.get_sharding_spec_by_name('m2')
+        output_sharding_spec = strategy.get_sharding_spec_by_name('addmm')
+        bias_sharding_spec = strategy.get_sharding_spec_by_name('input_1')
+
+        # make sure the sharding matches across different operation data
+        assert input_sharding_spec.sharding_sequence[:-1] == output_sharding_spec.sharding_sequence[:-1]
+        assert weight_sharding_spec.sharding_sequence[0] == input_sharding_spec.sharding_sequence[1]
+        assert weight_sharding_spec.sharding_sequence[1] == output_sharding_spec.sharding_sequence[1]
+        assert bias_sharding_spec.sharding_sequence[-1] == output_sharding_spec.sharding_sequence[-1]
+
+
+@parameterize('input_shape', [(16,), (4, 16)])
+@run_on_environment_flag(name='AUTO_PARALLEL')
+@pytest.mark.dist
+@rerun_if_address_is_in_use()
+def test_addmm_handler(input_shape):
+    world_size = 4
+    run_func_function = partial(check_linear_function_handler,
+                                input_shape=input_shape,
+                                world_size=world_size,
+                                port=free_port())
+    mp.spawn(run_func_function, nprocs=world_size)
+
+
+if __name__ == '__main__':
+    test_addmm_handler()