[refactor] remove gpc dependency in colotensor's _ops (#1189)

colossalai/nn/_ops/_utils.py

 import torch
 from typing import Union, Optional
 from colossalai.tensor import ColoTensor
+import torch
+import torch.distributed as dist
+from colossalai.global_variables import tensor_parallel_env as env
+
+from colossalai.nn.layer.utils import divide
+from colossalai.tensor import ProcessGroup
 
 GeneralTensor = Union[ColoTensor, torch.Tensor]
 Number = Union[int, float]
@@ -10,3 +16,182 @@ def convert_to_colo_tensor(tensor: Optional[GeneralTensor]) -> Optional[ColoTens
     if tensor is not None and not isinstance(tensor, ColoTensor):
         tensor = ColoTensor.from_torch_tensor(tensor)
     return tensor
+
+
+def set_parallel_input(input_parallel: bool):
+    env.parallel_input_1d = input_parallel
+
+
+def get_parallel_input():
+    return env.parallel_input_1d
+
+
+def vocab_range_from_per_partition_vocab_size(per_partition_vocab_size, rank):
+    index_f = rank * per_partition_vocab_size
+    index_l = index_f + per_partition_vocab_size
+    return index_f, index_l
+
+
+def vocab_range_from_global_vocab_size(global_vocab_size, rank, world_size):
+    per_partition_vocab_size = divide(global_vocab_size, world_size)
+    return vocab_range_from_per_partition_vocab_size(per_partition_vocab_size, rank)
+
+
+def _reduce(input_, pg: ProcessGroup):
+    # skip if only one rank involved
+    if pg.tp_world_size() == 1:
+        return input_
+    assert input_.device.type == 'cuda'
+    group = pg.tp_process_group()
+    dist.all_reduce(input_, group=group)
+
+    return input_
+
+
+def _split(input_, pg: ProcessGroup, dim=-1):
+    # skip if only one rank involved
+    world_size = pg.tp_world_size()
+    if world_size == 1:
+        return input_
+
+    # Split along last dimension.
+    dim_size = input_.size(dim)
+    assert dim_size % world_size == 0, \
+        f'The dimension to split ({dim_size}) is not a multiple of world size ({world_size}), ' \
+        f'cannot split tensor evenly'
+
+    tensor_list = torch.split(input_, dim_size // world_size, dim=dim)
+    rank = pg.tp_local_rank()
+    output = tensor_list[rank].contiguous()
+
+    return output
+
+
+def _gather(input_, pg: ProcessGroup, dim=-1):
+    # skip if only one rank involved
+    world_size = pg.tp_world_size()
+    if world_size == 1:
+        return input_
+
+    # all gather
+    rank = pg.tp_local_rank()
+    tensor_list = [torch.empty_like(input_) for _ in range(world_size)]
+    tensor_list[rank] = input_
+    assert input_.device.type == 'cuda'
+    group = pg.tp_process_group()
+    torch.distributed.all_gather(tensor_list, input_, group=group)
+
+    # concat
+    output = torch.cat(tensor_list, dim=dim).contiguous()
+
+    return output
+
+
+class _ReduceGrad(torch.autograd.Function):
+    """
+    Pass the input to the model parallel region.
+
+    Args:
+        input_: input matrix.
+        process_group: parallel mode.
+    """
+
+    @staticmethod
+    def symbolic(graph, input_):
+        return input_
+
+    @staticmethod
+    def forward(ctx, input_, process_group):
+        ctx.mode = process_group
+        return input_
+
+    @staticmethod
+    def backward(ctx, grad_output):
+        return _reduce(grad_output, ctx.mode), None
+
+
+class _ReduceInput(torch.autograd.Function):
+    """
+    All-reduce the input from the model parallel region.
+
+    Args:
+        input_: input matrix.
+        process_group: parallel mode.
+    """
+
+    @staticmethod
+    def symbolic(graph, input_):
+        return _reduce(input_)
+
+    @staticmethod
+    def forward(ctx, input_, process_group):
+        return _reduce(input_, process_group)
+
+    @staticmethod
+    def backward(ctx, grad_output):
+        return grad_output, None
+
+
+class _SplitForwardGatherBackward(torch.autograd.Function):
+    """
+    Split the input and keep only the corresponding chuck to the rank.
+    
+    Args:
+        input_: input matrix.
+        process_group: parallel mode.
+        dim: dimension
+    """
+
+    @staticmethod
+    def symbolic(graph, input_):
+        return _split(input_)
+
+    @staticmethod
+    def forward(ctx, input_, process_group, dim):
+        ctx.mode = process_group
+        ctx.dim = dim
+        return _split(input_, process_group, dim)
+
+    @staticmethod
+    def backward(ctx, grad_output):
+        return _gather(grad_output, ctx.mode, ctx.dim), None, None
+
+
+class _GatherForwardSplitBackward(torch.autograd.Function):
+    """Gather the input from model parallel region and concatenate.
+
+    Args:
+        input_: input matrix.
+        process_group: parallel mode.
+        dim: dimension
+    """
+
+    @staticmethod
+    def symbolic(graph, input_):
+        return _gather(input_)
+
+    @staticmethod
+    def forward(ctx, input_, process_group, dim):
+        ctx.mode = process_group
+        ctx.dim = dim
+        return _gather(input_, process_group, dim)
+
+    @staticmethod
+    def backward(ctx, grad_output):
+        return _split(grad_output, ctx.mode, ctx.dim), None, None
+
+
+def reduce_grad(input_, process_group):
+    return _ReduceGrad.apply(input_, process_group)
+
+
+def reduce_input(input_, process_group):
+    return _ReduceInput.apply(input_, process_group)
+
+
+def split_forward_gather_backward(input_, process_group, dim):
+    return _SplitForwardGatherBackward.apply(input_, process_group, dim)
+
+
+def gather_forward_split_backward(input_, process_group, dim):
+    return _GatherForwardSplitBackward.apply(input_, process_group, dim)

colossalai/nn/_ops/addmm.py

 import torch
 from colossalai.tensor.op_wrapper import colo_op_impl
-from colossalai.nn.layer.parallel_1d._utils import reduce_input, reduce_grad
 from colossalai.tensor import ComputePattern, TensorSpec, ComputePattern, ComputeSpec, ColoTensor
 from colossalai.tensor import distspec
-from colossalai.context import ParallelMode
 from ._utils import GeneralTensor, Number, convert_to_colo_tensor
+from ._utils import reduce_input, reduce_grad
 
 
 def colo_addmm_1Drow(input_tensor: ColoTensor, mat1: ColoTensor, mat2: ColoTensor, beta: Number,
@@ -12,18 +11,16 @@ def colo_addmm_1Drow(input_tensor: ColoTensor, mat1: ColoTensor, mat2: ColoTenso
     # mat1:S[1] x mat2:S[0] = Output:P
     # beta * input + alpha * All-Reduce(Output) = res
 
-    mat1 = mat1.convert_to_dist_spec(
-        distspec.shard(mat2.tensor_spec.get_process_group(), [-1], [mat2.tensor_spec.get_process_group_size()]))
+    mat1 = mat1.convert_to_dist_spec(distspec.shard(mat2.get_process_group(), [-1], [mat2.get_tp_world_size()]))
 
     # Output:P
     partial_output = torch.mm(mat1, mat2)
     # Reduce(Output)
-    output = reduce_input(partial_output, ParallelMode.PARALLEL_1D)
+    output = reduce_input(partial_output, mat1.get_process_group())
     # input
     assert not input_tensor.has_compute_spec(), 'Invalid input spec for 1Drow addmm op'
     output = beta * input_tensor + alpha * output
-    output = ColoTensor.from_torch_tensor(output,
-                                          spec=TensorSpec(distspec.replicate(mat2.tensor_spec.get_process_group())))
+    output = ColoTensor.from_torch_tensor(output, spec=TensorSpec(distspec.replicate(mat2.get_process_group())))
     return output
 
 
@@ -31,13 +28,12 @@ def colo_addmm_1Dcol(input_tensor: ColoTensor, mat1: ColoTensor, mat2: ColoTenso
                      alpha: Number) -> ColoTensor:
     # mat1:B x mat2:S[1] + input:S[1] = Output:S[1]
     compute_spec = mat2.tensor_spec.compute_spec
-    mat1 = mat1.convert_to_dist_spec(distspec.replicate(mat2.tensor_spec.get_process_group()))
-    mat1 = reduce_grad(mat1, ParallelMode.PARALLEL_1D)
+    mat1 = mat1.convert_to_dist_spec(distspec.replicate(mat2.get_process_group()))
+    mat1 = reduce_grad(mat1, mat1.get_process_group())
 
     output_parallel = torch.addmm(input_tensor, mat1, mat2, beta=beta, alpha=alpha)
-    output_spec = TensorSpec(
-        distspec.shard(mat2.tensor_spec.get_process_group(), [-1], [mat2.tensor_spec.get_process_group_size()]),
-        ComputeSpec(ComputePattern.TP1D))
+    output_spec = TensorSpec(distspec.shard(mat2.get_process_group(), [-1], [mat2.get_tp_world_size()]),
+                             ComputeSpec(ComputePattern.TP1D))
     output = ColoTensor.from_torch_tensor(output_parallel, spec=output_spec)
 
     if compute_spec.output_replicate:

colossalai/nn/_ops/embedding.py

 import torch.nn.functional as F
 from typing import Optional
 from colossalai.tensor.op_wrapper import colo_op_impl
-from colossalai.nn.layer.parallel_1d._utils import reduce_input
-from colossalai.core import global_context as gpc
 from colossalai.tensor import ComputePattern, TensorSpec, ComputePattern, ComputeSpec, ColoTensor, distspec
-from colossalai.context import ParallelMode
-from ._utils import GeneralTensor, convert_to_colo_tensor
+from ._utils import GeneralTensor, convert_to_colo_tensor, reduce_input
 
 
 def colo_embedding_1Dcol(input_tensor: ColoTensor,
@@ -17,7 +14,7 @@ def colo_embedding_1Dcol(input_tensor: ColoTensor,
                          sparse: bool = False) -> ColoTensor:
     # embedding_1Dcol split the weight(lookup table) to (num_embeddings, embedding_dim/P)
     # Gather splitted lookup table
-    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(weight.tensor_spec.get_process_group()))
+    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(weight.get_process_group()))
 
     output_parallel = F.embedding(input_tensor,
                                   weight,
@@ -26,9 +23,8 @@ def colo_embedding_1Dcol(input_tensor: ColoTensor,
                                   norm_type=norm_type,
                                   scale_grad_by_freq=scale_grad_by_freq,
                                   sparse=sparse)
-    output_spec = TensorSpec(
-        distspec.shard(weight.tensor_spec.get_process_group(), [-1], [weight.tensor_spec.get_process_group_size()]),
-        ComputeSpec(ComputePattern.TP1D))
+    output_spec = TensorSpec(distspec.shard(weight.get_process_group(), [-1], [weight.get_tp_world_size()]),
+                             ComputeSpec(ComputePattern.TP1D))
     output = ColoTensor.from_torch_tensor(output_parallel, spec=output_spec)
 
     compute_spec = weight.tensor_spec.compute_spec
@@ -49,9 +45,10 @@ def colo_embedding_1Drow(input_tensor: ColoTensor,
     # embedding_1Drow split the weight(lookup table) to (num_embeddings/P, embedding_dim)
     # Find index in this shard and mask those not here
     # Reduce all
-    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(weight.tensor_spec.get_process_group()))
+    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(weight.get_process_group()))
 
-    tensor_parallel_rank = gpc.get_local_rank(ParallelMode.PARALLEL_1D)
+    # tensor_parallel_rank = gpc.get_local_rank(ParallelMode.PARALLEL_1D)
+    tensor_parallel_rank = weight.tensor_spec.dist_spec.process_group.tp_local_rank()
     num_embeddings_per_partition = weight.size_local(0)
     vocab_start_index = tensor_parallel_rank * num_embeddings_per_partition
     vocab_end_index = vocab_start_index + num_embeddings_per_partition
@@ -75,9 +72,8 @@ def colo_embedding_1Drow(input_tensor: ColoTensor,
     # Mask the output embedding.
     partial_output[input_mask, :] = 0.
     # Reduce across all the model parallel GPUs.
-    output = reduce_input(partial_output, ParallelMode.PARALLEL_1D)
-    output = ColoTensor.from_torch_tensor(output,
-                                          spec=TensorSpec(distspec.replicate(weight.tensor_spec.get_process_group())))
+    output = reduce_input(partial_output, weight.get_process_group())
+    output = ColoTensor.from_torch_tensor(output, spec=TensorSpec(distspec.replicate(weight.get_process_group())))
     return output
 
 

colossalai/nn/_ops/embedding_bag.py

@@ -32,9 +32,8 @@ def colo_embedding_bag_1Dcol(input_tensor: ColoTensor,
                                       per_sample_weights=per_sample_weights,
                                       include_last_offset=include_last_offset,
                                       padding_idx=padding_idx)
-    output_spec = TensorSpec(
-        distspec.shard(weight.tensor_spec.get_process_group(), [-1], [weight.tensor_spec.get_process_group_size()]),
-        ComputeSpec(ComputePattern.TP1D))
+    output_spec = TensorSpec(distspec.shard(weight.get_process_group(), [-1], [weight.get_tp_world_size()]),
+                             ComputeSpec(ComputePattern.TP1D))
     output = ColoTensor.from_torch_tensor(output_parallel, spec=output_spec)
 
     if weight.tensor_spec.compute_spec.output_replicate:

colossalai/nn/_ops/layernorm.py

@@ -17,7 +17,7 @@ def colo_layernorm(
     input_tensor, weight, bias = tuple(map(convert_to_colo_tensor, (input_tensor, weight, bias)))
 
     # TODO (ver217): check dist spec
-    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(input_tensor.tensor_spec.get_process_group()))
+    input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(input_tensor.get_process_group()))
 
     output = F.layer_norm(input_tensor, normalized_shape, weight=weight, bias=bias, eps=eps)
     output = ColoTensor.from_torch_tensor(output, input_tensor.tensor_spec)

colossalai/nn/_ops/linear.py

@@ -2,9 +2,8 @@ import torch.nn.functional as F
 from typing import Optional
 from ._utils import GeneralTensor, convert_to_colo_tensor
 from colossalai.tensor.op_wrapper import colo_op_impl
-from colossalai.nn.layer.parallel_1d._utils import reduce_input, reduce_grad
+from ._utils import reduce_input, reduce_grad
 from colossalai.tensor import ComputePattern, TensorSpec, ComputePattern, ComputeSpec, ColoTensor, distspec
-from colossalai.context import ParallelMode
 from colossalai.nn.graph import register_colo_graph, GraphOpNode, GraphGlobalEnv
 
 
@@ -13,19 +12,18 @@ def colo_linear_1Drow(input_tensor: ColoTensor, weight: ColoTensor, bias: Option
     # All-Reduce(Output) + bias = res
     # Input:S[1]
     input_tensor = input_tensor.convert_to_dist_spec(
-        distspec.shard(weight.tensor_spec.get_process_group(), [-1], [weight.tensor_spec.get_process_group_size()]))
+        distspec.shard(weight.get_process_group(), [-1], [weight.get_tp_world_size()]))
 
     # Output:P
     partial_output = F.linear(input_tensor, weight)
     # Reduce(Output)
-    output = reduce_input(partial_output, ParallelMode.PARALLEL_1D)
+    output = reduce_input(partial_output, weight.get_process_group())
     # Bias
     if bias is not None:
         assert not bias.has_compute_spec(), 'Invalid bias spec for 1Drow Linear op'
         output = output + bias
 
-    output = ColoTensor.from_torch_tensor(output,
-                                          spec=TensorSpec(distspec.replicate(weight.tensor_spec.get_process_group())))
+    output = ColoTensor.from_torch_tensor(output, spec=TensorSpec(distspec.replicate(weight.get_process_group())))
     return output
 
 
@@ -35,13 +33,13 @@ def colo_linear_1Dcol(input_tensor: ColoTensor, weight: ColoTensor, bias: Option
     # Input:B
     compute_spec = weight.tensor_spec.compute_spec
     input_tensor = input_tensor.convert_to_dist_spec(distspec.replicate(weight.tensor_spec.get_process_group()))
-    input_parallel = reduce_grad(input_tensor, ParallelMode.PARALLEL_1D)
+    input_parallel = reduce_grad(input_tensor, weight.tensor_spec.dist_spec.process_group)
 
     output_parallel = F.linear(input_parallel, weight, bias)
     output = ColoTensor.from_torch_tensor(output_parallel,
                                           spec=TensorSpec(
-                                              distspec.shard(weight.tensor_spec.get_process_group(), [-1],
-                                                             [weight.tensor_spec.get_process_group_size()]),
+                                              distspec.shard(weight.get_process_group(), [-1],
+                                                             [weight.get_tp_world_size()]),
                                               ComputeSpec(ComputePattern.TP1D)))
     if compute_spec.output_replicate:
         return output.to_replicate()

colossalai/nn/parallel/data_parallel.py

 import torch
 import itertools
 import torch.distributed as dist
-from colossalai.core import global_context as gpc
-from colossalai.context import ParallelMode
 from functools import partial
 from colossalai.zero.utils.zero_hook_v2 import ZeROHookV2
 from colossalai.gemini.chunk import TensorState, Chunk
@@ -12,6 +10,7 @@ from typing import Dict, Iterable, List, Optional
 from colossalai.logging import get_dist_logger
 from collections import OrderedDict
 from colossalai.tensor.colo_parameter import ColoParameter
+from colossalai.tensor import ProcessGroup as ColoProcessGroup
 from .reducer import Reducer
 try:
     from torch.nn.modules.module import _EXTRA_STATE_KEY_SUFFIX, _IncompatibleKeys
@@ -45,8 +44,8 @@ class ColoDDP(torch.nn.Module):
         >>> from colossalai.core import global_context as gpc
         >>> from colossalai.context import ParallelMode
         >>> model = torch.nn.Linear(20, 1)
-        >>> model = ColoDDP(model)
-        >>> // model = ColoDDP(model, process_group=gpc.get_group(ParallelMode.DATA), cpu_process_group=gpc.get_cpu_group(ParallelMode.DATA))
+        >>> pg = ProcessGroup(tp_degree = world_size//2)
+        >>> model = ColoDDP(model, pg)
         >>> logits = model(x)
         >>> loss = criterion(logits, labels)
         >>> model.backward(loss)
@@ -55,13 +54,13 @@ class ColoDDP(torch.nn.Module):
         module (torch.nn.Module): Module to apply DDP.
         process_group (Optional[dist.ProcessGroup], optional): The process group which DDP uses.
             If it's None, the default data parallel group will be used. Defaults to None.
-        process_group (Optional[dist.ProcessGroup], optional): The process group which DDP uses for those parameters on CPU.
+        cpu_process_group (Optional[dist.ProcessGroup], optional): The process group which DDP uses for those parameters on CPU.
             If it's None, the default CPU data parallel group will be used. Defaults to None.
     """
 
     def __init__(self,
                  module: torch.nn.Module,
-                 process_group: Optional[dist.ProcessGroup] = None,
+                 process_group: ColoProcessGroup,
                  cpu_process_group: Optional[dist.ProcessGroup] = None,
                  bucket_cap_mb: int = 25,
                  rebuild_bucket: bool = True) -> None:
@@ -69,8 +68,9 @@ class ColoDDP(torch.nn.Module):
         super().__init__()
         self.module = module
         self.comm_stream: torch.cuda.Stream = torch.cuda.Stream()
-        self.process_group = process_group or gpc.get_group(ParallelMode.DATA)
-        self.cpu_process_group = cpu_process_group or gpc.get_cpu_group(ParallelMode.DATA)
+        assert process_group
+
+        self.process_group = process_group.dp_process_group()
         self.dp_world_size = self.process_group.size()
         self.reducer = Reducer(bucket_cap_mb)
         self.rebuild_bucket = rebuild_bucket
@@ -120,6 +120,8 @@ class ColoDDP(torch.nn.Module):
             return empty_grad
 
         else:
+            #TODO(jiaruifang) fixme
+            raise NotImplementedError
             dist.all_reduce(grad, group=self.cpu_process_group)
             return grad
 
@@ -191,8 +193,11 @@ class ZeroDDP(ColoDDP):
             For more details, see the API reference of ``GeminiManager``.
     """
 
-    def __init__(self, module: torch.nn.Module, gemini_manager: GeminiManager) -> None:
-        super().__init__(module.half())
+    def __init__(self,
+                 module: torch.nn.Module,
+                 gemini_manager: GeminiManager,
+                 process_group: Optional[ColoProcessGroup] = None) -> None:
+        super().__init__(module.half(), process_group=process_group)
         self.gemini_manager = gemini_manager
         self.chunk_manager = gemini_manager.chunk_manager
         self.param_op_hook = ZeROHookV2(gemini_manager)

colossalai/nn/parallel/layers/colo_module.py

@@ -52,5 +52,5 @@ class ColoModule(object):
     def get_param_names(self):
         return self._shard_params
 
-    def register(self, compute_pattern):
+    def register(self, compute_pattern, pg):
         raise NotImplementedError

colossalai/nn/parallel/layers/embedding.py

 from .colo_module import ColoModule
-from colossalai.tensor import ComputePattern, distspec
+from colossalai.tensor import ComputePattern, distspec, ProcessGroup
 from colossalai.core import global_context as gpc
 from colossalai.context.parallel_mode import ParallelMode
 
@@ -10,20 +10,18 @@ class ColoEmbedding(ColoModule):
         super(ColoEmbedding, self).__init__()
         self._register_shard_params(['weight'])
 
-    def register(self, compute_pattern):
+    def register(self, compute_pattern, pg: ProcessGroup):
         if not compute_pattern in self._allowed_patterns:
             if ComputePattern.TP1D == compute_pattern:
-                self._set_TP1D()
+                self._set_TP1D(pg)
 
-    def _set_TP1D(self):
+    def _set_TP1D(self, pg: ProcessGroup):
         # TP1D Row Linear
         _compute_pattern = ComputePattern.TP1D
         self._register_allowed_patterns(
             compute_pattern=_compute_pattern,
             dist_specs={
-                'weight':
-                    distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [0],
-                                   [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
+                'weight': distspec.shard(pg, [0], [pg.tp_world_size()]),
             },
             mode='row',
         )
@@ -32,9 +30,7 @@ class ColoEmbedding(ColoModule):
         self._register_allowed_patterns(
             compute_pattern=_compute_pattern,
             dist_specs={
-                'weight':
-                    distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [-1],
-                                   [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
+                'weight': distspec.shard(pg, [-1], [pg.tp_world_size()]),
             },
             mode='col',
         )

colossalai/nn/parallel/layers/linear.py

 from .colo_module import ColoModule
-from colossalai.tensor import ComputePattern, distspec
-from colossalai.core import global_context as gpc
-from colossalai.context.parallel_mode import ParallelMode
+from colossalai.tensor import ComputePattern, distspec, ProcessGroup
 
 
 class ColoLinear(ColoModule):
@@ -10,22 +8,19 @@ class ColoLinear(ColoModule):
         super(ColoLinear, self).__init__()
         self._register_shard_params(['weight', 'bias'])
 
-    def register(self, compute_pattern):
+    def register(self, compute_pattern, pg: ProcessGroup):
         if not compute_pattern in self._allowed_patterns:
             if ComputePattern.TP1D == compute_pattern:
-                self._set_TP1D()
+                self._set_TP1D(pg)
 
-    def _set_TP1D(self):
+    def _set_TP1D(self, pg):
         # TP1D Row Linear
         _compute_pattern = ComputePattern.TP1D
         self._register_allowed_patterns(
             compute_pattern=_compute_pattern,
             dist_specs={
-                'weight':
-                    distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [-1],
-                                   [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
-                'bias':
-                    None
+                'weight': distspec.shard(pg, [-1], [pg.tp_world_size()]),
+                'bias': None
             },
             mode='row',
         )
@@ -34,12 +29,8 @@ class ColoLinear(ColoModule):
         self._register_allowed_patterns(
             compute_pattern=_compute_pattern,
             dist_specs={
-                'weight':
-                    distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [0],
-                                   [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
-                'bias':
-                    distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [0],
-                                   [gpc.get_world_size(ParallelMode.PARALLEL_1D)])
+                'weight': distspec.shard(pg, [0], [pg.tp_world_size()]),
+                'bias': distspec.shard(pg, [0], [pg.tp_world_size()])
             },
             mode='col',
         )

colossalai/nn/parallel/layers/module_utils.py

 from typing import Dict
-from colossalai.tensor import ColoParameter, ComputeSpec, TensorSpec
+from colossalai.tensor import ColoParameter, ComputeSpec, TensorSpec, ProcessGroup
 from . import ColoModule
 import torch
 
@@ -29,7 +29,7 @@ def get_colo_module(module: torch.nn.Module):
         return None
 
 
-def check_colo_module(module: torch.nn.Module, recursive=True):
+def check_colo_module(module: torch.nn.Module, pg: ProcessGroup, recursive=True):
     if is_colo_module(module):
         colo_module = get_colo_module(module)
         param_names = colo_module.get_param_names()
@@ -50,7 +50,7 @@ def check_colo_module(module: torch.nn.Module, recursive=True):
                 continue
 
         if compute_pattern is not None:
-            colo_module.register(compute_pattern)
+            colo_module.register(compute_pattern, pg)
             if not colo_module.has_compute_pattern(compute_pattern):
                 raise Exception(
                     f'Invalid ColoParameter spec: ComputePattern {compute_pattern} in {module} is not allowed.')
@@ -76,16 +76,20 @@ def check_colo_module(module: torch.nn.Module, recursive=True):
                 raise Exception(f'Invalid ColoParameter spec: Params in {module} are incorrectly sharded.')
     if recursive == True:
         for submodule in module.children():
-            check_colo_module(submodule, recursive=True)
+            check_colo_module(submodule, pg=pg, recursive=True)
 
 
-def init_colo_module(module: torch.nn.Module, compute_spec: ComputeSpec, recursive=True, mode='default'):
+def init_colo_module(module: torch.nn.Module,
+                     compute_spec: ComputeSpec,
+                     pg: ProcessGroup,
+                     recursive=True,
+                     mode='default'):
     compute_pattern = compute_spec.compute_pattern
     if is_colo_module(module):
         # for each param
         # set DistSpec and ComputeSpec
         colo_module = get_colo_module(module)
-        colo_module.register(compute_pattern)
+        colo_module.register(compute_pattern, pg)
         if not colo_module.has_compute_pattern_with_mode(compute_pattern, mode=mode):
             raise NotImplementedError
         # a set for modules which update at least one param in the init process.
@@ -101,7 +105,7 @@ def init_colo_module(module: torch.nn.Module, compute_spec: ComputeSpec, recursi
                 for mod in param.shared_param_modules:
                     modules_update_param.add(mod)
         for mod in modules_update_param:
-            check_colo_module(mod, recursive=False)
+            check_colo_module(mod, pg, recursive=False)
     if recursive == True:
         for submodule in module.children():
-            init_colo_module(submodule, compute_spec, recursive=True, mode=mode)
+            init_colo_module(submodule, compute_spec, pg=pg, recursive=True, mode=mode)

colossalai/tensor/colo_tensor.py

@@ -78,6 +78,12 @@ class ColoTensor(torch.Tensor):
     def is_model_data(self) -> bool:
         return self._type == TensorType.MODEL
 
+    def get_process_group(self) -> 'ProcessGroup':
+        return self._tensor_spec.dist_spec.process_group
+
+    def get_tp_world_size(self) -> int:
+        return self._tensor_spec.dist_spec.process_group.tp_world_size()
+
     @classmethod
     def __torch_function__(cls, func, types, args=(), kwargs=None):
         if kwargs is None:

colossalai/tensor/dist_spec_mgr.py

@@ -5,6 +5,7 @@ from contextlib import contextmanager
 import torch
 import torch.distributed as dist
 from packaging import version
+from colossalai.logging import get_dist_logger
 
 
 # TODO(jiaruifang) circle import, move the divide to colossalai.commons.
@@ -64,7 +65,7 @@ class DistSpecManager:
         DistSpecManager._sanity_check(old_dist_spec, dist_spec)
 
         chunk = tensor
-        idx = dist_spec.process_group.rank()
+        idx = dist_spec.process_group.tp_local_rank()
         num_parts = prod(dist_spec.num_partitions)
         for i, dim in enumerate(dist_spec.dims):
             num_parts //= dist_spec.num_partitions[i]
@@ -91,8 +92,9 @@ class DistSpecManager:
             saved_dev = tensor.device
             tensor.data = tensor.data.cuda()
 
-        buffer = [torch.empty_like(tensor) for _ in range(old_dist_spec.process_group.size())]
-        dist.all_gather(buffer, tensor, group=old_dist_spec.process_group)
+        buffer = [torch.empty_like(tensor) for _ in range(old_dist_spec.process_group.tp_world_size())]
+        assert tensor.device.type == 'cuda'
+        dist.all_gather(buffer, tensor, group=old_dist_spec.process_group.tp_process_group())
         for i in range(len(old_dist_spec.dims) - 1, -1, -1):
             new_buffer = []
             dim = old_dist_spec.dims[i]
@@ -108,14 +110,14 @@ class DistSpecManager:
 
     @staticmethod
     def _all_to_all(tensor: torch.Tensor, old_dist_spec: _DistSpec, dist_spec: _DistSpec) -> torch.Tensor:
-        world_size = old_dist_spec.process_group.size()
+        world_size = old_dist_spec.process_group.tp_world_size()
         if world_size == 1:
             return tensor
 
-        assert tensor.device.type == "cuda" and dist.get_backend(old_dist_spec.process_group) == "nccl", \
+        assert tensor.device.type == "cuda" and old_dist_spec.process_group.backend == "nccl", \
             "Currently, only CUDA Tensor with NCCL backend is supported for the requested AlltoAll " \
             f"collective function, however, we got {tensor.device.type} device and " \
-            f"{dist.get_backend(old_dist_spec.process_group)} backend"
+            f"{old_dist_spec.process_group.backend} backend"
 
         gather_dim = old_dist_spec.dims[0]
         scatter_dim = dist_spec.dims[0]
@@ -126,7 +128,7 @@ class DistSpecManager:
 
         scatter_list = [t.contiguous() for t in torch.tensor_split(tensor, world_size, scatter_dim)]
         gather_list = [torch.empty(*shapes, dtype=tensor.dtype, device=tensor.device) for _ in range(world_size)]
-        dist.all_to_all(gather_list, scatter_list, group=old_dist_spec.process_group)
+        dist.all_to_all(gather_list, scatter_list, group=old_dist_spec.process_group.tp_process_group())
 
         output_ = torch.cat(gather_list, dim=gather_dim).contiguous()
         assert output_.shape[scatter_dim] == scattered_dim_size and output_.shape[gather_dim] == gathered_dim_size

colossalai/tensor/distspec.py

 from enum import Enum
-from torch.distributed import ProcessGroup
+from colossalai.tensor import ProcessGroup
 from typing import Optional, List
 from numpy import prod
 
@@ -51,8 +51,8 @@ def replicate(process_group: Optional[ProcessGroup] = None) -> _DistSpec:
 
 
 def shard(process_group: ProcessGroup, dims: List[int], num_partitions: List[int]) -> _DistSpec:
-    assert process_group is not None
+    assert process_group is not None and isinstance(process_group, ProcessGroup)
     assert isinstance(dims, list) and isinstance(num_partitions, list)
     assert len(dims) == len(num_partitions)
-    assert prod(num_partitions) == process_group.size(), f"{num_partitions} {process_group.size()}"
+    assert prod(num_partitions) == process_group.tp_world_size(), f"{num_partitions} {process_group.tp_world_size()}"
     return _DistSpec(DistPlacementPattern.SHARD, process_group, dims=tuple(dims), num_partitions=tuple(num_partitions))

colossalai/tensor/process_group.py

 import torch
 from typing import List, Optional
+from colossalai.logging import get_dist_logger
 
 
 class ProcessGroup:
@@ -41,12 +42,12 @@ class ProcessGroup:
         if dp_degree and not tp_degree:
             self._dp_degree = dp_degree
             assert self._world_size % self._dp_degree == 0, f"DP degree {dp_degree} should be divisible by {self._world_size} hen DP degree is None"
-            self._tp_degree = self._world_size / dp_degree
+            self._tp_degree = self._world_size // dp_degree
 
         if not dp_degree and tp_degree:
             self._tp_degree = tp_degree
             assert self._world_size % self._tp_degree == 0, f"TP degree {tp_degree} should be divisible by {self._world_size} when DP degree is None"
-            self._dp_degree = self._world_size / tp_degree
+            self._dp_degree = self._world_size // tp_degree
 
         self._tp_rank_list = []
         self._dp_rank_list = []
@@ -58,12 +59,48 @@ class ProcessGroup:
             if rank_id // self._tp_degree == self._rank // self._tp_degree:
                 self._tp_rank_list.append(rank_id)
 
-        self._tp_process_group = torch.distributed.new_group(ranks=self._tp_rank_list, backend=backend)
-        self._dp_process_group = torch.distributed.new_group(ranks=self._dp_rank_list, backend=backend)
+        assert backend == 'nccl'
+        self._tp_process_group = torch.distributed.new_group(ranks=self._tp_rank_list)
+        self._dp_process_group = torch.distributed.new_group(ranks=self._dp_rank_list)
+
+        self.logger = get_dist_logger('ProcessGroup')
+        self.logger.info(f'{self._rank} initialize TP group on {self._tp_rank_list} DP group pn {self._dp_rank_list}')
+
+    @property
+    def backend(self):
+        return self._backend
+
+    def __eq__(self, obj: 'ProcessGroup') -> bool:
+        if not isinstance(obj, ProcessGroup):
+            return False
+        if self._rank != obj._rank:
+            assert False
+        if self._rank_list != obj._rank_list:
+            assert False
+        if self._tp_rank_list != obj._tp_rank_list:
+            assert False
+        if self._dp_rank_list != obj._dp_rank_list:
+            assert False
+        if self._backend != obj._backend:
+            assert False
+        if self._tp_degree != obj._tp_degree:
+            return False
+        if self._dp_degree != obj._dp_degree:
+            return False
+        return True
+
+    def rank(self):
+        return self._rank
 
     def world_size(self):
         return self._world_size
 
+    def tp_local_rank(self):
+        return self._rank % self._tp_degree
+
+    def dp_local_rank(self):
+        return self._rank // self._tp_degree
+
     def dp_world_size(self):
         return len(self._dp_rank_list)
 

colossalai/tensor/tensor_spec.py

@@ -17,11 +17,12 @@ class TensorSpec(object):
         self.compute_spec = compute_spec
         self.dist_spec = dist_spec
 
+    # TODO(jiaruifang) actually need tp process group
     def get_process_group(self):
         return self.dist_spec.process_group
 
     def get_process_group_size(self):
-        return dist.get_world_size(self.dist_spec.process_group)
+        return dist.get_world_size(self.dist_spec.process_group.tp_process_group())
 
     def get_placement(self):
         return self.dist_spec.placement
@@ -30,7 +31,7 @@ class TensorSpec(object):
         return self.dist_spec.placement == DistPlacementPattern.REPLICATE \
             or (len(self.dist_spec.num_partitions) == 1
                 and self.dist_spec.num_partitions[0] == 1) \
-            or (self.dist_spec.process_group.size() == 1)
+            or (self.dist_spec.process_group.tp_world_size() == 1)
 
     def is_shard_1dcol(self):
         return self.dist_spec.placement == DistPlacementPattern.SHARD \

tests/test_ddp/test_ddp_ignore_params.py

@@ -15,6 +15,7 @@ import torch.distributed as dist
 import os
 import random
 import numpy as np
+from colossalai.tensor import ProcessGroup
 
 
 def set_seed(seed):
@@ -27,14 +28,16 @@ def set_seed(seed):
 
 
 def init_ddp(module: torch.nn.Module) -> ColoDDP:
-    return ColoDDP(module)
+    pg = ProcessGroup()
+    return ColoDDP(module, process_group=pg)
 
 
 def init_ddpv2(module: torch.nn.Module, use_chunk: bool = False) -> ZeroDDP:
     chunk_size = ChunkManager.search_chunk_size(module, 64, 2) if use_chunk else None
     chunk_manager = ChunkManager(chunk_size)
     gemini_manager = GeminiManager('cuda', chunk_manager)
-    return ZeroDDP(module, gemini_manager)
+    pg = ProcessGroup()
+    return ZeroDDP(module, gemini_manager, pg)
 
 
 class Net(torch.nn.Module):

tests/test_ddp/test_ddp_state_dict.py

@@ -13,6 +13,7 @@ from colossalai.nn.parallel import ZeroDDP, ColoDDP
 from colossalai.gemini.gemini_mgr import GeminiManager
 from typing import Callable
 from collections import OrderedDict
+from colossalai.tensor import ProcessGroup
 
 
 def check_state_dict_equal(state_dict: OrderedDict, other_state_dict: OrderedDict):
@@ -22,14 +23,16 @@ def check_state_dict_equal(state_dict: OrderedDict, other_state_dict: OrderedDic
 
 
 def init_ddp(module: torch.nn.Module) -> ColoDDP:
-    return ColoDDP(module)
+    pg = ProcessGroup()
+    return ColoDDP(module, process_group=pg)
 
 
 def init_ddpv2(module: torch.nn.Module, use_chunk: bool = False, use_zero: bool = False) -> ZeroDDP:
     chunk_size = ChunkManager.search_chunk_size(module, 64, 4) if use_chunk else None
     chunk_manager = ChunkManager(chunk_size, enable_distributed_storage=use_zero)
     gemini_manager = GeminiManager('cuda', chunk_manager)
-    return ZeroDDP(module, gemini_manager)
+    pg = ProcessGroup()
+    return ZeroDDP(module, gemini_manager, process_group=pg)
 
 
 def run_state_dict(ddp_init_func: Callable[[torch.nn.Module], ColoDDP]):

tests/test_tensor/_utils/_util.py

@@ -41,7 +41,7 @@ def tensor_equal(A, B):
     return torch.allclose(A, B, rtol=1e-3, atol=1e-1)
 
 
-def tensor_shard_equal(tensor: torch.Tensor, shard: torch.Tensor):
+def tensor_shard_equal(tensor: torch.Tensor, shard: torch.Tensor, rank, world_size):
     assert tensor.ndim == shard.ndim
     if tensor.shape == shard.shape:
         return tensor_equal(tensor, shard)
@@ -50,8 +50,10 @@ def tensor_shard_equal(tensor: torch.Tensor, shard: torch.Tensor):
         if dims_not_eq.numel() == 1:
             # 1D shard
             dim = dims_not_eq.item()
-            world_size = gpc.get_world_size(ParallelMode.PARALLEL_1D)
-            rank = gpc.get_local_rank(ParallelMode.PARALLEL_1D)
+            if world_size is None:
+                world_size = gpc.get_world_size(ParallelMode.PARALLEL_1D)
+            if rank is None:
+                rank = gpc.get_local_rank(ParallelMode.PARALLEL_1D)
             return tensor_equal(tensor.chunk(world_size, dim)[rank], shard)
         else:
             raise NotImplementedError

tests/test_tensor/test_addmm_tp.py

@@ -3,14 +3,12 @@ import torch
 import pytest
 import torch.nn as nn
 import torch.multiprocessing as mp
-from colossalai.tensor import ColoTensor
+from colossalai.tensor import ColoTensor, ProcessGroup
 from colossalai.tensor import distspec
 from colossalai.tensor import TensorSpec, ComputePattern, ComputeSpec, DistSpecManager
-from colossalai.context import ParallelMode
 from colossalai.testing import rerun_if_address_is_in_use
 from colossalai.utils import free_port
 from functools import partial
-from colossalai.core import global_context as gpc
 from _utils import tensor_shard_equal, tensor_equal
 
 
@@ -38,18 +36,14 @@ class Conv1D(nn.Module):
         return x
 
 
-def init_1d_row(weight, bias):
-    spec = TensorSpec(
-        distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [0], [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
-        ComputeSpec(ComputePattern.TP1D))
+def init_1d_row(weight, bias, pg: ProcessGroup):
+    spec = TensorSpec(distspec.shard(pg, [0], [pg.tp_world_size()]), ComputeSpec(ComputePattern.TP1D))
     with DistSpecManager.no_grad():
         weight.set_tensor_spec(spec)
 
 
-def init_1d_col(weight, bias):
-    spec = TensorSpec(
-        distspec.shard(gpc.get_group(ParallelMode.PARALLEL_1D), [-1], [gpc.get_world_size(ParallelMode.PARALLEL_1D)]),
-        ComputeSpec(ComputePattern.TP1D))
+def init_1d_col(weight, bias, pg: ProcessGroup):
+    spec = TensorSpec(distspec.shard(pg, [-1], [pg.tp_world_size()]), ComputeSpec(ComputePattern.TP1D))
     with DistSpecManager.no_grad():
         weight.set_tensor_spec(spec)
         bias.set_tensor_spec(spec)
@@ -59,7 +53,9 @@ def run_with_spec(spec_init_func):
     model = Conv1D(4, 16).cuda()
     weight = ColoTensor(torch.nn.Parameter(model.weight.detach()))
     bias = ColoTensor(torch.nn.Parameter(model.bias.detach()))
-    spec_init_func(weight, bias)
+    world_size = torch.distributed.get_world_size()
+    pg = ProcessGroup(tp_degree=world_size)
+    spec_init_func(weight, bias, pg)
     x = torch.rand(2, 16).cuda()
     out = model(x)
     colo_out = torch.addmm(bias, x, weight)
@@ -68,13 +64,12 @@ def run_with_spec(spec_init_func):
     grad = torch.rand_like(out)
     out.backward(grad)
     colo_out.backward(grad)
-    tensor_shard_equal(model.weight.grad, weight.grad)
-    tensor_shard_equal(model.bias.grad, bias.grad)
+    tensor_shard_equal(model.weight.grad, weight.grad, pg.tp_local_rank(), pg.tp_world_size())
+    tensor_shard_equal(model.bias.grad, bias.grad, pg.tp_local_rank(), pg.tp_world_size())
 
 
 def run_dist(rank, world_size, port):
-    config = dict(parallel=dict(tensor=dict(mode="1d", size=world_size),))
-    colossalai.launch(config=config, rank=rank, world_size=world_size, host='localhost', port=port, backend='nccl')
+    colossalai.launch(config={}, rank=rank, world_size=world_size, host='localhost', port=port, backend='nccl')
     run_with_spec(init_1d_row)
     run_with_spec(init_1d_col)