2022-12-02 10:13:20 +00:00
|
|
|
from typing import Dict
|
2023-09-19 06:20:26 +00:00
|
|
|
|
|
|
|
import torch
|
2022-07-12 06:10:37 +00:00
|
|
|
from torch.fx.graph import Graph
|
2023-09-19 06:20:26 +00:00
|
|
|
from torch.fx.node import Node, map_arg
|
|
|
|
|
2022-07-07 08:22:00 +00:00
|
|
|
|
|
|
|
def get_comm_size(prev_partition, next_partition):
|
2022-07-12 06:10:37 +00:00
|
|
|
"""
|
|
|
|
Given two partitions (parent and child),
|
2022-07-07 08:22:00 +00:00
|
|
|
calculate the communication size between the two.
|
|
|
|
"""
|
|
|
|
# Keep tracking the communication size between parent and child
|
|
|
|
comm_size = 0
|
|
|
|
# Keep tracking all the counted node
|
|
|
|
visited_nodes = set()
|
|
|
|
# Go through all nodes in the child partition
|
|
|
|
# If a node has input nodes from the parent partition,
|
|
|
|
# the output size of those input nodes will be counted
|
|
|
|
# and added to comm_size
|
2022-07-08 02:54:41 +00:00
|
|
|
parent_node_names = [n.name for n in prev_partition.graph.nodes]
|
|
|
|
for node in next_partition.graph.nodes:
|
2022-07-07 08:22:00 +00:00
|
|
|
input_nodes: Dict[Node, None] = {}
|
|
|
|
map_arg(node.args, lambda n: input_nodes.setdefault(n))
|
|
|
|
map_arg(node.kwargs, lambda n: input_nodes.setdefault(n))
|
|
|
|
for n in input_nodes:
|
|
|
|
if n.name in parent_node_names and n not in visited_nodes:
|
2023-09-19 06:20:26 +00:00
|
|
|
comm_size += n.meta["tensor_meta"].numel
|
2022-07-07 08:22:00 +00:00
|
|
|
visited_nodes.add(n)
|
|
|
|
return comm_size
|
2022-07-12 06:10:37 +00:00
|
|
|
|
|
|
|
|
|
|
|
def get_leaf(graph: Graph):
|
|
|
|
"""
|
|
|
|
Given a graph, return leaf nodes of this graph.
|
|
|
|
Note: If we remove ``root`` nodes, ``placeholder`` nodes, and ``output`` nodes from fx graph,
|
|
|
|
we will get a normal DAG. Leaf nodes in this context means leaf nodes in that DAG.
|
|
|
|
"""
|
|
|
|
input_nodes: Dict[Node, None] = {}
|
|
|
|
for node in graph.nodes:
|
2023-09-19 06:20:26 +00:00
|
|
|
if node.op == "output":
|
2022-07-12 06:10:37 +00:00
|
|
|
map_arg(node.args, lambda n: input_nodes.setdefault(n))
|
|
|
|
map_arg(node.kwargs, lambda n: input_nodes.setdefault(n))
|
|
|
|
placeholder_nodes = []
|
|
|
|
for node in input_nodes.keys():
|
2023-09-19 06:20:26 +00:00
|
|
|
if node.op == "placeholder":
|
2022-07-12 06:10:37 +00:00
|
|
|
placeholder_nodes.append(node)
|
|
|
|
for node in placeholder_nodes:
|
|
|
|
input_nodes.pop(node)
|
|
|
|
return list(input_nodes.keys())
|
|
|
|
|
|
|
|
|
|
|
|
def is_leaf(graph: Graph, node: Node):
|
|
|
|
return node in get_leaf(graph)
|
|
|
|
|
|
|
|
|
|
|
|
def get_top(graph: Graph):
|
|
|
|
"""
|
|
|
|
Given a graph, return top nodes of this graph.
|
|
|
|
Note: If we remove ``root`` nodes, ``placeholder`` nodes, and ``output`` nodes from fx graph,
|
|
|
|
we will get a normal DAG. Top nodes in this context means nodes with BFS level 0 in that DAG.
|
|
|
|
"""
|
|
|
|
top_node_list = set()
|
|
|
|
for node in graph.nodes:
|
2023-09-19 06:20:26 +00:00
|
|
|
if node.op == "output":
|
2022-07-12 06:10:37 +00:00
|
|
|
continue
|
|
|
|
is_top = False
|
|
|
|
|
|
|
|
def _get_top(node):
|
|
|
|
nonlocal is_top
|
2023-09-19 06:20:26 +00:00
|
|
|
if node.op == "placeholder":
|
2022-07-12 06:10:37 +00:00
|
|
|
is_top = True
|
|
|
|
|
|
|
|
map_arg(node.args, lambda n: _get_top(n))
|
|
|
|
map_arg(node.kwargs, lambda n: _get_top(n))
|
|
|
|
if is_top:
|
|
|
|
top_node_list.add(node)
|
|
|
|
return list(top_node_list)
|
|
|
|
|
|
|
|
|
|
|
|
def is_top(graph: Graph, node: Node):
|
|
|
|
return node in get_top(graph)
|
|
|
|
|
|
|
|
|
|
|
|
def get_all_consumers(graph: Graph, node: Node):
|
|
|
|
"""
|
|
|
|
Given a graph and a node of this graph, return all consumers of the node.
|
2023-09-19 06:20:26 +00:00
|
|
|
|
2022-07-12 06:10:37 +00:00
|
|
|
Returns:
|
|
|
|
List of ``Nodes`` that node appear in these nodes ``args`` and ``kwargs``.
|
|
|
|
"""
|
|
|
|
consumer_list = []
|
|
|
|
for n in graph.nodes:
|
|
|
|
if node in n.all_input_nodes:
|
|
|
|
consumer_list.append(n)
|
|
|
|
return consumer_list
|
|
|
|
|
|
|
|
|
|
|
|
def assign_bfs_level_to_nodes(graph: Graph):
|
|
|
|
"""
|
|
|
|
Give a graph, assign bfs level to each node of this graph excluding ``placeholder`` and ``output`` nodes.
|
|
|
|
Example:
|
|
|
|
class MLP(torch.nn.Module):
|
|
|
|
def __init__(self, dim: int):
|
|
|
|
super().__init__()
|
|
|
|
self.linear1 = torch.nn.Linear(dim, dim)
|
|
|
|
self.linear2 = torch.nn.Linear(dim, dim)
|
|
|
|
self.linear3 = torch.nn.Linear(dim, dim)
|
|
|
|
self.linear4 = torch.nn.Linear(dim, dim)
|
|
|
|
self.linear5 = torch.nn.Linear(dim, dim)
|
|
|
|
def forward(self, x):
|
|
|
|
l1 = self.linear1(x)
|
|
|
|
l2 = self.linear2(x)
|
|
|
|
l3 = self.linear3(l1)
|
|
|
|
l4 = self.linear4(l2)
|
|
|
|
l5 = self.linear5(l3)
|
|
|
|
return l4, l5
|
|
|
|
model = MLP(4)
|
|
|
|
gm = symbolic_trace(model)
|
|
|
|
print(gm.graph)
|
|
|
|
assign_bfs_level_to_nodes(gm.graph)
|
|
|
|
for node in gm.graph.nodes:
|
|
|
|
if hasattr(node, 'bfs_level'):
|
|
|
|
print(node.name, node.bfs_level)
|
2023-09-19 06:20:26 +00:00
|
|
|
|
2022-07-12 06:10:37 +00:00
|
|
|
Output:
|
|
|
|
graph():
|
|
|
|
%x : [#users=2] = placeholder[target=x]
|
|
|
|
%linear1 : [#users=1] = call_module[target=linear1](args = (%x,), kwargs = {})
|
|
|
|
%linear2 : [#users=1] = call_module[target=linear2](args = (%x,), kwargs = {})
|
|
|
|
%linear3 : [#users=1] = call_module[target=linear3](args = (%linear1,), kwargs = {})
|
|
|
|
%linear4 : [#users=1] = call_module[target=linear4](args = (%linear2,), kwargs = {})
|
|
|
|
%linear5 : [#users=1] = call_module[target=linear5](args = (%linear3,), kwargs = {})
|
|
|
|
return (linear4, linear5)
|
|
|
|
linear1 0
|
|
|
|
linear2 0
|
|
|
|
linear3 1
|
|
|
|
linear4 1
|
|
|
|
linear5 2
|
|
|
|
"""
|
|
|
|
current_level = 0
|
|
|
|
nodes_to_process = []
|
|
|
|
|
|
|
|
top_nodes = get_top(graph)
|
|
|
|
for node in top_nodes:
|
|
|
|
node.bfs_level = current_level
|
|
|
|
nodes_to_process.extend(get_all_consumers(graph, node))
|
|
|
|
|
|
|
|
current_level += 1
|
|
|
|
while nodes_to_process:
|
|
|
|
new_process_list = []
|
|
|
|
for node in nodes_to_process:
|
2023-09-19 06:20:26 +00:00
|
|
|
if node.op == "output":
|
2022-07-12 06:10:37 +00:00
|
|
|
continue
|
|
|
|
node.bfs_level = current_level
|
|
|
|
new_process_list.extend(get_all_consumers(graph, node))
|
|
|
|
nodes_to_process = new_process_list
|
|
|
|
current_level += 1
|
2022-08-30 07:54:37 +00:00
|
|
|
|
|
|
|
|
|
|
|
def get_node_module(node) -> torch.nn.Module:
|
|
|
|
"""
|
|
|
|
Find the module associated with the given node.
|
|
|
|
Args:
|
|
|
|
node (torch.fx.Node): a torch.fx.Node object in the fx computation graph
|
|
|
|
Returns:
|
|
|
|
torch.nn.Module: the module associated with the given node
|
|
|
|
"""
|
|
|
|
|
2023-09-19 06:20:26 +00:00
|
|
|
assert (
|
|
|
|
node.graph.owning_module is not None
|
|
|
|
), "Cannot find the owning_module for node.graph, please make sure the graph is associated with a GraphModule object"
|
|
|
|
assert node.op == "call_module", f"Expected node.op to be call_module, but found {node.op}"
|
2022-08-30 07:54:37 +00:00
|
|
|
module = node.graph.owning_module.get_submodule(node.target)
|
|
|
|
return module
|