You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/colossalai/autochunk/autochunk_codegen.py

616 lines
25 KiB

from typing import Any, Callable, Dict, Iterable, List, Tuple
import torch
import colossalai
from colossalai.fx._compatibility import is_compatible_with_meta
2 years ago
from colossalai.fx.codegen.activation_checkpoint_codegen import CODEGEN_AVAILABLE
AUTOCHUNK_AVAILABLE = CODEGEN_AVAILABLE and is_compatible_with_meta()
if AUTOCHUNK_AVAILABLE:
from torch.fx.graph import (
CodeGen,
PythonCode,
_custom_builtins,
_CustomBuiltin,
_format_target,
_is_from_torch,
_Namespace,
_origin_type_map,
inplace_methods,
magic_methods,
)
2 years ago
from torch.fx.node import Argument, Node, _get_qualified_name, _type_repr, map_arg
2 years ago
2 years ago
from .search_chunk import SearchChunk
from .utils import delete_free_var_from_last_use, get_logger, get_node_name, get_node_shape
2 years ago
def _gen_chunk_slice_dim(chunk_dim: int, chunk_indice_name: str, shape: List) -> str:
"""
Generate chunk slice string, eg. [:, :, chunk_idx_name:chunk_idx_name + chunk_size, :]
Args:
chunk_dim (int)
chunk_indice_name (str): chunk indice name
shape (List): node shape
Returns:
new_shape (str): return slice
"""
new_shape = "["
2 years ago
for idx, _ in enumerate(shape):
if idx == chunk_dim:
2 years ago
new_shape += "%s:%s + chunk_size" % (chunk_indice_name, chunk_indice_name)
else:
new_shape += ":"
new_shape += ", "
new_shape = new_shape[:-2] + "]"
return new_shape
def _gen_loop_start(chunk_input: List[Node], chunk_output: List[Node], chunk_output_dim: int, chunk_size=2) -> str:
2 years ago
"""
Generate chunk loop start
eg. chunk_result = torch.empty([100, 100], dtype=input_node.dtype, device=input_node.device)
chunk_size = 32
for chunk_idx in range(0, 100, 32):
......
Args:
chunk_input (List[Node]): chunk input node
chunk_output (Node): chunk output node
chunk_output_dim (int): chunk output node chunk dim
2 years ago
chunk_size (int): chunk size. Defaults to 2.
Returns:
context (str): generated str
"""
input_node = chunk_input[0]
context = ""
for i in range(len(chunk_output)):
shape_str = str(list(get_node_shape(chunk_output[i])))
if get_node_name(chunk_output[i]) in ["split", "unbind"]:
tensor_str = "torch.empty(%s, dtype=%s.dtype, device=%s.device), " % (
shape_str,
input_node.name,
input_node.name,
)
tensor_str = tensor_str * len(chunk_output[i].meta["tensor_meta"])
tensor_str = "[" + tensor_str[:-2] + "]"
context += "%s = %s; " % (chunk_output[i].name, tensor_str)
else:
context += "%s = torch.empty(%s, dtype=%s.dtype, device=%s.device); " % (
chunk_output[i].name,
shape_str,
input_node.name,
input_node.name,
)
out_shape = get_node_shape(chunk_output[0])
chunk_shape = out_shape[chunk_output_dim[0]]
context += "chunk_size = %d\nfor chunk_idx in range(0, %d, chunk_size):\n" % (chunk_size, chunk_shape)
return context
def _gen_loop_end(
chunk_inputs: List[Node],
chunk_non_compute_inputs: List[Node],
node_list: List[Node],
chunk_outputs_idx: int,
chunk_outputs_non_tensor: List[Node],
search_chunk: SearchChunk,
) -> str:
2 years ago
"""
Generate chunk loop end
eg. chunk_result[chunk_idx:chunk_idx + chunk_size] = output_node
output_node = chunk_result; xx = None; xx = None
Args:
chunk_inputs (List[Node]): chunk input node
chunk_non_compute_inputs (List[Node]): input node without chunk
chunk_outputs (Node): chunk output node
chunk_outputs_dim (int): chunk output node chunk dim
node_list (List)
Returns:
context (str): generated str
2 years ago
"""
context = "chunk_size = None"
# determine if its the last use for chunk input
for chunk_input in chunk_inputs + chunk_non_compute_inputs:
if all([search_chunk.node_mgr.find_node_idx(user) <= chunk_outputs_idx for user in chunk_input.users.keys()]):
context += "; %s = None" % chunk_input.name
for chunk_output_non_tensor, chunk_output_non_tensor_val in chunk_outputs_non_tensor.items():
context += "; %s = %s" % (chunk_output_non_tensor.name, chunk_output_non_tensor_val)
context += "\n"
return context
2 years ago
def _replace_name(context: str, name_from: str, name_to: str) -> str:
"""
replace node name
"""
patterns = [(" ", " "), (" ", "."), (" ", ","), ("(", ")"), ("(", ","), (" ", ")"), (" ", ""), ("", " ")]
for p in patterns:
source = p[0] + name_from + p[1]
target = p[0] + name_to + p[1]
if source in context:
context = context.replace(source, target)
break
return context
2 years ago
def _replace_reshape_size(context: str, node_name: str, reshape_size_dict: Dict) -> str:
"""
replace reshape size, some may have changed due to chunk
"""
if node_name not in reshape_size_dict:
return context
context = context.replace(reshape_size_dict[node_name][0], reshape_size_dict[node_name][1])
return context
def _replace_new_tensor_like_shape(
2 years ago
search_chunk: SearchChunk,
chunk_infos: List[Dict],
region_idx: int,
node_idx: int,
node: Node,
body: List[str],
) -> List[str]:
"""
add chunk slice for new tensor op such as ones like
"""
if get_node_name(node) in ["ones_like", "zeros_like", "empty_like"]:
meta_node = search_chunk.node_mgr.get_node_by_idx(node_idx)
2 years ago
chunk_dim = chunk_infos[region_idx]["node_chunk_dim"][meta_node]["chunk_dim"]
if get_node_shape(meta_node)[chunk_dim] != 1:
source_node = meta_node.args[0].args[0]
if (
source_node not in chunk_infos[region_idx]["node_chunk_dim"]
or chunk_infos[region_idx]["node_chunk_dim"][source_node]["chunk_dim"] is None
):
chunk_slice = _gen_chunk_slice_dim(chunk_dim, "chunk_idx", get_node_shape(node))
body[-1] = _replace_name(body[-1], node.args[0].name, node.args[0].name + chunk_slice)
2 years ago
return body
def _replace_new_tensor_shape(
search_chunk: SearchChunk,
chunk_infos: List[Dict],
region_idx: int,
node_idx: int,
node: Node,
body: List[str],
) -> List[str]:
"""
add chunk slice for new tensor op such as ones
"""
if get_node_name(node) in ["ones", "zeros", "empty"]:
meta_node = search_chunk.node_mgr.get_node_by_idx(node_idx)
chunk_dim = chunk_infos[region_idx]["node_chunk_dim"][meta_node]["chunk_dim"]
if chunk_dim is None:
return
if get_node_shape(meta_node)[chunk_dim] == 1:
return
origin_shape = str(node.args)
new_shape = list(node.args)
new_shape[chunk_dim] = "min(chunk_size, %d - chunk_idx)" % get_node_shape(meta_node)[chunk_dim]
new_shape = str(new_shape)
new_shape = new_shape.replace("'", "")
body[-1] = _replace_name(body[-1], origin_shape[1:-1], new_shape[1:-1])
return body
def _add_node_slice(
chunk_nodes: List[Node],
2 years ago
region_idx: int,
chunk_nodes_dim: Dict,
2 years ago
node_idx: int,
body: List[str],
node: Node,
2 years ago
) -> List[str]:
"""
add chunk slice for input nodes
"""
for chunk_node_idx, chunk_node in enumerate(chunk_nodes[region_idx]):
# inputs node
if isinstance(chunk_nodes_dim[region_idx][chunk_node_idx], dict):
for idx, dim in chunk_nodes_dim[region_idx][chunk_node_idx].items():
if idx == node_idx:
chunk_slice = _gen_chunk_slice_dim(dim[0], "chunk_idx", get_node_shape(chunk_node))
body[-1] = _replace_name(body[-1], chunk_node.name, chunk_node.name + chunk_slice)
# outputs node
else:
if chunk_node.name == node.name or (chunk_node.name in [i.name for i in node.all_input_nodes]):
chunk_slice = _gen_chunk_slice_dim(
chunk_nodes_dim[region_idx][chunk_node_idx], "chunk_idx", get_node_shape(chunk_node)
)
if get_node_name(chunk_node) in ["split", "unbind"]:
split_chunk_slice = ""
for i in range(len(chunk_node.meta["tensor_meta"])):
split_chunk_slice += "%s[%d]%s, " % (chunk_node.name, i, chunk_slice)
split_chunk_slice = split_chunk_slice[:-2]
body[-1] = _replace_name(body[-1], chunk_node.name, split_chunk_slice)
else:
body[-1] = _replace_name(body[-1], chunk_node.name, chunk_node.name + chunk_slice)
2 years ago
return body
def emit_code_with_chunk(
body: List[str],
nodes: Iterable[Node],
emit_node_func: Callable,
delete_unused_value_func: Callable,
search_chunk: SearchChunk,
chunk_infos: List,
eval_mem: bool = False,
):
2 years ago
"""
Emit code with chunk according to chunk_infos.
2 years ago
2 years ago
It will generate a for loop in chunk regions, and
2 years ago
replace inputs and outputs of regions with chunked variables.
Args:
body: forward code
nodes: graph.nodes
emit_node_func: function to emit node
delete_unused_value_func: function to remove the unused value
2 years ago
search_chunk: the class to search all chunks
chunk_infos: store all information about all chunks.
"""
node_list = list(nodes)
2 years ago
# chunk region
chunk_starts = [i["region"][0] for i in chunk_infos]
chunk_ends = [i["region"][1] for i in chunk_infos]
2 years ago
# chunk inputs
chunk_inputs = [i["inputs"] for i in chunk_infos] # input with chunk
chunk_inputs_non_chunk = [i["inputs_non_chunk"] for i in chunk_infos] # input without chunk
chunk_inputs_dim = [i["inputs_dim"] for i in chunk_infos] # input chunk dim
chunk_inputs_names = [j.name for i in chunk_inputs for j in i] + [j.name for i in chunk_inputs_non_chunk for j in i]
2 years ago
# chunk outputs
chunk_outputs = [i["outputs"] for i in chunk_infos]
chunk_outputs_non_tensor = [i["outputs_non_tensor"] for i in chunk_infos]
chunk_outputs_dim = [i["outputs_dim"] for i in chunk_infos]
2 years ago
node_list = search_chunk.reorder_graph.reorder_node_list(node_list)
node_idx = 0
region_idx = 0
within_chunk_region = False
if eval_mem:
body.append("init_memory = torch.cuda.memory_allocated() / 1024**2\n")
while node_idx < len(node_list):
node = node_list[node_idx]
2 years ago
# if is chunk start, generate for loop start
if node_idx in chunk_starts:
within_chunk_region = True
region_idx = chunk_starts.index(node_idx)
body.append(
_gen_loop_start(
chunk_inputs[region_idx],
chunk_outputs[region_idx],
chunk_outputs_dim[region_idx],
chunk_infos[region_idx]["chunk_size"],
)
)
if within_chunk_region:
emit_node_func(node, body)
# replace input var with chunk var
body = _add_node_slice(chunk_inputs, region_idx, chunk_inputs_dim, node_idx, body, node)
# replace output var with chunk var
body = _add_node_slice(chunk_outputs, region_idx, chunk_outputs_dim, node_idx, body, node)
# new tensor like
body = _replace_new_tensor_like_shape(search_chunk, chunk_infos, region_idx, node_idx, node, body)
# new tensor
body = _replace_new_tensor_shape(search_chunk, chunk_infos, region_idx, node_idx, node, body)
# reassign reshape size
body[-1] = _replace_reshape_size(body[-1], node.name, chunk_infos[region_idx]["reshape_size"])
body[-1] = " " + body[-1]
delete_unused_value_func(node, body, chunk_inputs_names)
if eval_mem:
body.append(
" if chunk_idx == 0:\n print('%s', torch.cuda.max_memory_allocated() / 1024**2 - init_memory); torch.cuda.reset_peak_memory_stats()\n"
% (node.name)
)
else:
emit_node_func(node, body)
if node_idx not in chunk_inputs:
delete_unused_value_func(node, body, chunk_inputs_names)
if eval_mem:
body.append(
"print('%s', torch.cuda.max_memory_allocated() / 1024**2 - init_memory); torch.cuda.reset_peak_memory_stats()\n"
% (node.name)
)
2 years ago
# generate chunk region end
if node_idx in chunk_ends:
body.append(
_gen_loop_end(
chunk_inputs[region_idx],
chunk_inputs_non_chunk[region_idx],
node_list,
chunk_ends[region_idx],
chunk_outputs_non_tensor[region_idx],
search_chunk,
)
)
within_chunk_region = False
node_idx += 1
if AUTOCHUNK_AVAILABLE:
2 years ago
class AutoChunkCodeGen(CodeGen):
def __init__(
self,
meta_graph,
max_memory: int = None,
print_mem: bool = False,
print_progress: bool = False,
eval_mem: bool = False,
) -> None:
2 years ago
super().__init__()
self.eval_mem = eval_mem
2 years ago
# find the chunk regions
self.search_chunk = SearchChunk(meta_graph, max_memory, print_mem, print_progress)
2 years ago
self.chunk_infos = self.search_chunk.search_region()
if print_progress:
get_logger().info("AutoChunk start codegen")
def _gen_python_code(self, nodes, root_module: str, namespace: _Namespace, verbose=None) -> PythonCode:
2 years ago
free_vars: List[str] = []
body: List[str] = []
globals_: Dict[str, Any] = {}
wrapped_fns: Dict[str, None] = {}
2 years ago
# Wrap string in list to pass by reference
maybe_return_annotation: List[str] = [""]
2 years ago
def add_global(name_hint: str, obj: Any):
"""Add an obj to be tracked as a global.
2 years ago
We call this for names that reference objects external to the
Graph, like functions or types.
Returns: the global name that should be used to reference 'obj' in generated source.
"""
if _is_from_torch(obj) and obj != torch.device: # to support registering torch.device
2 years ago
# HACK: workaround for how torch custom ops are registered. We
# can't import them like normal modules so they must retain their
# fully qualified name.
return _get_qualified_name(obj)
# normalize the name hint to get a proper identifier
global_name = namespace.create_name(name_hint, obj)
if global_name in globals_:
assert globals_[global_name] is obj
return global_name
globals_[global_name] = obj
return global_name
2 years ago
# set _custom_builtins here so that we needn't import colossalai in forward
_custom_builtins["colossalai"] = _CustomBuiltin("import colossalai", colossalai)
2 years ago
# Pre-fill the globals table with registered builtins.
for name, (_, obj) in _custom_builtins.items():
add_global(name, obj)
2 years ago
def type_repr(o: Any):
if o == ():
# Empty tuple is used for empty tuple type annotation Tuple[()]
return "()"
2 years ago
typename = _type_repr(o)
2 years ago
if hasattr(o, "__origin__"):
# This is a generic type, e.g. typing.List[torch.Tensor]
origin_type = _origin_type_map.get(o.__origin__, o.__origin__)
origin_typename = add_global(_type_repr(origin_type), origin_type)
2 years ago
if hasattr(o, "__args__"):
# Assign global names for each of the inner type variables.
args = [type_repr(arg) for arg in o.__args__]
2 years ago
if len(args) == 0:
# Bare type, such as `typing.Tuple` with no subscript
# This code-path used in Python < 3.9
return origin_typename
2 years ago
return f'{origin_typename}[{",".join(args)}]'
else:
# Bare type, such as `typing.Tuple` with no subscript
2 years ago
# This code-path used in Python 3.9+
return origin_typename
2 years ago
# Common case: this is a regular module name like 'foo.bar.baz'
return add_global(typename, o)
def _format_args(args: Tuple[Argument, ...], kwargs: Dict[str, Argument]) -> str:
2 years ago
def _get_repr(arg):
# Handle NamedTuples (if it has `_fields`) via add_global.
if isinstance(arg, tuple) and hasattr(arg, "_fields"):
qualified_name = _get_qualified_name(type(arg))
global_name = add_global(qualified_name, type(arg))
return f"{global_name}{repr(tuple(arg))}"
return repr(arg)
args_s = ", ".join(_get_repr(a) for a in args)
kwargs_s = ", ".join(f"{k} = {_get_repr(v)}" for k, v in kwargs.items())
if args_s and kwargs_s:
return f"{args_s}, {kwargs_s}"
return args_s or kwargs_s
# Run through reverse nodes and record the first instance of a use
# of a given node. This represents the *last* use of the node in the
# execution order of the program, which we will use to free unused
# values
node_to_last_use: Dict[Node, Node] = {}
user_to_last_uses: Dict[Node, List[Node]] = {}
def register_last_uses(n: Node, user: Node):
if n not in node_to_last_use:
node_to_last_use[n] = user
user_to_last_uses.setdefault(user, []).append(n)
for node in reversed(nodes):
map_arg(node.args, lambda n: register_last_uses(n, node))
map_arg(node.kwargs, lambda n: register_last_uses(n, node))
delete_free_var_from_last_use(user_to_last_uses)
# NOTE: we add a variable to distinguish body and ckpt_func
def delete_unused_values(user: Node, body, to_keep=[]):
"""
Delete values after their last use. This ensures that values that are
not used in the remainder of the code are freed and the memory usage
of the code is optimal.
"""
if user.op == "placeholder":
return
if user.op == "output":
body.append("\n")
return
nodes_to_delete = user_to_last_uses.get(user, [])
nodes_to_delete = [i for i in nodes_to_delete if i.name not in to_keep]
if len(nodes_to_delete):
to_delete_str = " = ".join([repr(n) for n in nodes_to_delete] + ["None"])
2 years ago
body.append(f"; {to_delete_str}\n")
else:
2 years ago
body.append("\n")
2 years ago
# NOTE: we add a variable to distinguish body and ckpt_func
def emit_node(node: Node, body):
maybe_type_annotation = "" if node.type is None else f" : {type_repr(node.type)}"
2 years ago
if node.op == "placeholder":
assert isinstance(node.target, str)
maybe_default_arg = "" if not node.args else f" = {repr(node.args[0])}"
free_vars.append(f"{node.target}{maybe_type_annotation}{maybe_default_arg}")
2 years ago
raw_name = node.target.replace("*", "")
if raw_name != repr(node):
body.append(f"{repr(node)} = {raw_name}\n")
return
elif node.op == "call_method":
assert isinstance(node.target, str)
body.append(
2 years ago
f"{repr(node)}{maybe_type_annotation} = {_format_target(repr(node.args[0]), node.target)}"
f"({_format_args(node.args[1:], node.kwargs)})"
)
return
2 years ago
elif node.op == "call_function":
assert callable(node.target)
# pretty print operators
if node.target.__module__ == "_operator" and node.target.__name__ in magic_methods:
2 years ago
assert isinstance(node.args, tuple)
body.append(
f"{repr(node)}{maybe_type_annotation} = "
f"{magic_methods[node.target.__name__].format(*(repr(a) for a in node.args))}"
)
2 years ago
return
# pretty print inplace operators; required for jit.script to work properly
# not currently supported in normal FX graphs, but generated by torchdynamo
if node.target.__module__ == "_operator" and node.target.__name__ in inplace_methods:
body.append(
f"{inplace_methods[node.target.__name__].format(*(repr(a) for a in node.args))}; "
f"{repr(node)}{maybe_type_annotation} = {repr(node.args[0])}"
)
2 years ago
return
qualified_name = _get_qualified_name(node.target)
global_name = add_global(qualified_name, node.target)
# special case for getattr: node.args could be 2-argument or 3-argument
# 2-argument: attribute access; 3-argument: fall through to attrib function call with default value
if (
global_name == "getattr"
and isinstance(node.args, tuple)
and isinstance(node.args[1], str)
and node.args[1].isidentifier()
and len(node.args) == 2
):
2 years ago
body.append(
f"{repr(node)}{maybe_type_annotation} = {_format_target(repr(node.args[0]), node.args[1])}"
)
2 years ago
return
body.append(
f"{repr(node)}{maybe_type_annotation} = {global_name}({_format_args(node.args, node.kwargs)})"
)
2 years ago
if node.meta.get("is_wrapped", False):
wrapped_fns.setdefault(global_name)
return
2 years ago
elif node.op == "call_module":
assert isinstance(node.target, str)
body.append(
f"{repr(node)}{maybe_type_annotation} = "
f"{_format_target(root_module, node.target)}({_format_args(node.args, node.kwargs)})"
)
return
2 years ago
elif node.op == "get_attr":
assert isinstance(node.target, str)
body.append(f"{repr(node)}{maybe_type_annotation} = {_format_target(root_module, node.target)}")
2 years ago
return
elif node.op == "output":
if node.type is not None:
maybe_return_annotation[0] = f" -> {type_repr(node.type)}"
body.append(self.generate_output(node.args[0]))
return
raise NotImplementedError(f"node: {node.op} {node.target}")
# Modified for activation checkpointing
ckpt_func = []
# if any node has a list of labels for activation_checkpoint, we
# will use nested type of activation checkpoint codegen
emit_code_with_chunk(
body, nodes, emit_node, delete_unused_values, self.search_chunk, self.chunk_infos, self.eval_mem
)
2 years ago
if len(body) == 0:
# If the Graph has no non-placeholder nodes, no lines for the body
# have been emitted. To continue to have valid Python code, emit a
# single pass statement
body.append("pass\n")
if len(wrapped_fns) > 0:
wrap_name = add_global("wrap", torch.fx.wrap)
wrap_stmts = "\n".join([f'{wrap_name}("{name}")' for name in wrapped_fns])
2 years ago
else:
wrap_stmts = ""
2 years ago
if self._body_transformer:
body = self._body_transformer(body)
2 years ago
for name, value in self.additional_globals():
add_global(name, value)
2 years ago
# as we need colossalai.utils.checkpoint, we need to import colossalai
# in forward function
prologue = self.gen_fn_def(free_vars, maybe_return_annotation[0])
prologue = "".join(ckpt_func) + prologue
prologue = prologue
2 years ago
code = "".join(body)
code = "\n".join(" " + line for line in code.split("\n"))
fn_code = f"""
2 years ago
{wrap_stmts}
2 years ago
{prologue}
{code}"""
2 years ago
# print(fn_code)
return PythonCode(fn_code, globals_)