[autoparallel] added node handler for bmm (#1655)

colossalai/auto_parallel/solver/op_handler/dot_handler_v2.py

@@ -2,11 +2,11 @@ import torch
 import torch.nn.functional as F
 from .node_handler import ModuleHandler, NodeHandler
 from ..sharding_strategy import ShardingStrategy_V2, OperationDataType, OperationData
-from ..strategy import LinearProjectionStrategyGenerator, StrategyGenerator_V2
+from ..strategy import LinearProjectionStrategyGenerator, StrategyGenerator_V2, BatchedMatMulStrategyGenerator
 from typing import List, Dict
 from .registry import operator_registry
 
-__all__ = ['LinearModuleHandler', 'LinearFunctionHandler']
+__all__ = ['LinearModuleHandler', 'LinearFunctionHandler', 'BMMFunctionHandler']
 
 
 @operator_registry.register(torch.nn.Linear)
@@ -133,3 +133,30 @@ class LinearFunctionHandler(NodeHandler):
                 # re-init the sharding spec
                 sharding_spec.__init__(sharding_spec.device_mesh, sharding_spec.entire_shape, dim_partition_dict)
         return strategy
+
+
+@operator_registry.register(torch.bmm)
+@operator_registry.register(torch.Tensor.bmm)
+class BMMFunctionHandler(NodeHandler):
+
+    def get_operation_data_mapping(self) -> Dict[str, OperationData]:
+        # use transposed shape for strategies
+        # the strategies will be transformed back to its original shape in self.post_process
+        physical_input_operand = OperationData(name=str(self.node.args[0]),
+                                               type=OperationDataType.ARG,
+                                               data=self.node.args[0]._meta_data)
+
+        physical_other_operand = OperationData(name=str(self.node.args[1]),
+                                               type=OperationDataType.ARG,
+                                               data=self.node.args[1]._meta_data)
+        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}
+        return mapping
+
+    def get_strategy_generator(self) -> List[StrategyGenerator_V2]:
+        generators = []
+        op_data_mapping = self.get_operation_data_mapping()
+        generators = []
+        generators.append(BatchedMatMulStrategyGenerator(op_data_mapping, self.device_mesh))
+        return generators

colossalai/auto_parallel/solver/strategy/matmul_strategy_generator.py

@@ -483,6 +483,9 @@ class BatchedMatMulStrategyGenerator(MatMulStrategyGenerator):
         other_op_data = self.op_data['other']
         assert input_op_data.data.dim() > 2 or other_op_data.data.dim() > 2
 
+    def update_compute_cost(self, strategy: ShardingStrategy_V2) -> ShardingStrategy_V2:
+        return self.op_data['input'].data.shape[-1] * reduce(operator.mul, self.op_data['output'].data.shape)
+
     def split_one_batch_dim(self):
         device_mesh_is_1d = True
         if len(self.device_mesh.mesh_shape) == 1:
@@ -552,7 +555,7 @@ class BatchedMatMulStrategyGenerator(MatMulStrategyGenerator):
             },
             "bias": {},
             "output": {
-                0: mesh_dim_0,
+                0: [mesh_dim_0],
                 -2: [mesh_dim_1]
             }
         }
@@ -635,8 +638,10 @@ class BatchedMatMulStrategyGenerator(MatMulStrategyGenerator):
         # can be None as it is only for 1D device mesh
         strategy = self.split_one_batch_dim()
         if strategy:
+            # only for 1D device mesh
             strategy_list.append(strategy)
-
+        else:
+            # for 2D device mesh
             # split batch dim of two inputs and the i dim of the first tensor
             # SbSi = SbSi x Sb
             strategy_list.append(self.split_batch_dim_lhs_space(0, 1))

colossalai/auto_parallel/solver/strategy/strategy_generator.py

@@ -49,6 +49,7 @@ class StrategyGenerator_V2(ABC):
         """
         results = {}
         for op_data_name, dim_partition_dict in mapping.items():
+            if op_data_name in self.op_data:
                 op_data = self.op_data[op_data_name]
                 sharding_spec = ShardingSpec(device_mesh=self.device_mesh,
                                              entire_shape=op_data.logical_shape,

tests/test_auto_parallel/test_node_handler/test_bmm_handler.py

+import pytest
+import torch
+import torch.nn as nn
+from colossalai.fx import ColoTracer, ColoGraphModule
+from colossalai.auto_parallel.solver.op_handler.dot_handler_v2 import BMMFunctionHandler
+from colossalai.auto_parallel.solver.sharding_strategy import OperationData, OperationDataType, StrategiesVector
+from colossalai.device.device_mesh import DeviceMesh
+
+
+class BMMTensorMethodModule(nn.Module):
+
+    def forward(self, x1, x2):
+        return x1.bmm(x2)
+
+
+class BMMTorchFunctionModule(nn.Module):
+
+    def forward(self, x1, x2):
+        return torch.bmm(x1, x2)
+
+
+@pytest.mark.parametrize('module', [BMMTensorMethodModule, BMMTorchFunctionModule])
+def test_2d_device_mesh(module):
+
+    model = module()
+    tracer = ColoTracer()
+    graph = tracer.trace(model,
+                         meta_args={
+                             "x1": torch.rand(4, 8, 16).to('meta'),
+                             'x2': torch.rand(4, 16, 8).to('meta')
+                         })
+    print(graph)
+    gm = ColoGraphModule(model, graph)
+    physical_mesh_id = torch.arange(0, 4)
+
+    mesh_shape = (2, 2)
+    device_mesh = DeviceMesh(physical_mesh_id, mesh_shape)
+    linear_mod_node = list(graph.nodes)[2]
+    strategies_vector = StrategiesVector(linear_mod_node)
+
+    # build handler
+    handler = BMMFunctionHandler(node=linear_mod_node, device_mesh=device_mesh, strategies_vector=strategies_vector)
+
+    # check operation data mapping
+    mapping = handler.get_operation_data_mapping()
+
+    for name, op_data in mapping.items():
+        op_data: OperationData
+        # make sure they have valid values
+        assert op_data.logical_shape is not None
+        assert op_data.data is not None
+
+    assert mapping['input'].name == "x1"
+    assert mapping['input'].data.is_meta
+    assert mapping['input'].data.shape == torch.Size([4, 8, 16])
+    assert mapping['input'].type == OperationDataType.ARG
+    assert mapping['input'].logical_shape == torch.Size([4, 8, 16])
+
+    assert mapping['other'].name == "x2"
+    assert mapping['other'].data.is_meta
+    assert mapping['other'].data.shape == torch.Size([4, 16, 8])
+    assert mapping['other'].type == OperationDataType.ARG
+    assert mapping['other'].logical_shape == torch.Size([4, 16, 8])
+
+    assert mapping['output'].name == "bmm"
+    assert mapping['output'].data.is_meta
+    assert mapping['output'].data.shape == torch.Size([4, 8, 8])
+    assert mapping['output'].type == OperationDataType.OUTPUT
+
+    strategies_vector = handler.register_strategy()
+    strategy_name_list = [val.name for val in strategies_vector]
+
+    # one batch dim
+    assert 'Sb0 = Sb0 x Sb0' not in strategy_name_list
+
+    # two batch dim
+    assert 'Sb01 = Sb01 x Sb01' in strategy_name_list
+
+    # SbSi = SbSi x Sb
+    assert 'Sb0Si1 = Sb0Si1 x Sb0' in strategy_name_list
+    assert 'Sb1Si0 = Sb1Si0 x Sb1' in strategy_name_list
+
+    # SbSj = SbR x SbSj
+    assert 'Sb0Sj1 = Sb0R x Sb0Sj1' in strategy_name_list
+    assert 'Sb1Sj0 = Sb1R x Sb1Sj0' in strategy_name_list
+
+    # SbR = SbSk x SbSk
+    assert 'Sb0R = Sb0Sk1 x Sb0Sk1' in strategy_name_list
+    assert 'Sb1R = Sb1Sk0 x Sb1Sk0' in strategy_name_list
+
+
+@pytest.mark.parametrize('module', [BMMTensorMethodModule, BMMTorchFunctionModule])
+def test_1d_device_mesh(module):
+    model = module()
+    tracer = ColoTracer()
+    graph = tracer.trace(model,
+                         meta_args={
+                             "x1": torch.rand(4, 8, 16).to('meta'),
+                             'x2': torch.rand(4, 16, 8).to('meta')
+                         })
+    print(graph)
+    gm = ColoGraphModule(model, graph)
+    physical_mesh_id = torch.arange(0, 4)
+
+    mesh_shape = (1, 4)
+    device_mesh = DeviceMesh(physical_mesh_id, mesh_shape)
+    linear_mod_node = list(graph.nodes)[2]
+    strategies_vector = StrategiesVector(linear_mod_node)
+
+    # build handler
+    handler = BMMFunctionHandler(node=linear_mod_node, device_mesh=device_mesh, strategies_vector=strategies_vector)
+
+    # check operation data mapping
+    mapping = handler.get_operation_data_mapping()
+
+    for name, op_data in mapping.items():
+        op_data: OperationData
+        # make sure they have valid values
+        assert op_data.logical_shape is not None
+        assert op_data.data is not None
+
+    assert mapping['input'].name == "x1"
+    assert mapping['input'].data.is_meta
+    assert mapping['input'].data.shape == torch.Size([4, 8, 16])
+    assert mapping['input'].type == OperationDataType.ARG
+    assert mapping['input'].logical_shape == torch.Size([4, 8, 16])
+
+    assert mapping['other'].name == "x2"
+    assert mapping['other'].data.is_meta
+    assert mapping['other'].data.shape == torch.Size([4, 16, 8])
+    assert mapping['other'].type == OperationDataType.ARG
+    assert mapping['other'].logical_shape == torch.Size([4, 16, 8])
+
+    assert mapping['output'].name == "bmm"
+    assert mapping['output'].data.is_meta
+    assert mapping['output'].data.shape == torch.Size([4, 8, 8])
+    assert mapping['output'].type == OperationDataType.OUTPUT
+
+    strategies_vector = handler.register_strategy()
+    strategy_name_list = [val.name for val in strategies_vector]
+    assert len(strategy_name_list) == 1
+    # one batch dim
+    assert 'Sb0 = Sb0 x Sb0' in strategy_name_list
+
+
+if __name__ == '__main__':
+    test_1d_device_mesh(BMMTensorMethodModule)
+    test_1d_device_mesh(BMMTorchFunctionModule)
+    test_2d_device_mesh(BMMTensorMethodModule)
+    test_2d_device_mesh(BMMTorchFunctionModule)