Readable Dump Documentation

Overview

Readable Dump is a high-readability graph structure export format provided by GE (Graph Engine), using text representation similar to Dynamo fx graph style, facilitating developers to directly read and understand computation graph structures.

Design Principles

Readability as highest priority, simplicity over completeness:

Display method: Through function-call-like approach, display IR attributes, inputs, outputs information
Information filtering: Non-IR defined information like control edges and private attributes not displayed (affects overall readability)
Value abbreviation: Values carried by Const use abbreviated display, see Appendix: Const Constant Display Rules

readable File Parsing

Example 1: Below is a complete ge_readable*.txt file:

graph("MakeTransformerSubGraph"):
  %input_0 : [#users=1] = Node[type=Data] (attrs = {index: 0})
  %input_1 : [#users=1] = Node[type=Data] (attrs = {index: 1})
  %input_2 : [#users=1] = Node[type=Data] (attrs = {index: 2})
  %Const_0 : [#users=1] = Node[type=Const] (attrs = {value: [-1 7168]})
  %Reshape_1 : [#users=1] = Node[type=Reshape] (inputs = (x=%input_0, shape=%Const_0), attrs = {axis: 0, num_axes: -1})
  %Cast_2 : [#users=1] = Node[type=Cast] (inputs = (x=%Reshape_1), attrs = {dst_type: 0})
  %Cast_3 : [#users=1] = Node[type=Cast] (inputs = (x=%input_1), attrs = {dst_type: 0})
  %Const_4 : [#users=1] = Node[type=Const] (attrs = {value: [1 0]})
  %Transpose_5 : [#users=1] = Node[type=Transpose] (inputs = (x=%Cast_3, perm=%Const_4))
  %MatMul_6 : [#users=1] = Node[type=MatMul] (inputs = (x1=%Cast_2, x2=%Transpose_5), attrs = {transpose_x1: false, transpose_x2: false})
  %Sigmoid_7 : [#users=1] = Node[type=Sigmoid] (inputs = (x=%MatMul_6))
  %Const_8 : [#users=1] = Node[type=Const] (attrs = {value: [-1 256]})
  %Reshape_9 : [#users=1] = Node[type=Reshape] (inputs = (x=%Sigmoid_7, shape=%Const_8), attrs = {axis: 0, num_axes: -1})
  %Unsqueeze_10 : [#users=1] = Node[type=Unsqueeze] (inputs = (x=%input_2), attrs = {axes: {0}})
  %Cast_11 : [#users=1] = Node[type=Cast] (inputs = (x=%Unsqueeze_10), attrs = {dst_type: 0})
  %Add_12 : [#users=1] = Node[type=Add] (inputs = (x1=%Reshape_9, x2=%Cast_11))
  %Const_13 : [#users=1] = Node[type=Const] (attrs = {value: [2]})
  %TopKV2_14 : [#users=2] = Node[type=TopKV2] (inputs = (x=%Add_12, k=%Const_13), attrs = {sorted: true, dim: -1, largest: true, indices_dtype: 3})
  %ret : [users=1] = get_element[node=%TopKV2_14](0)
  %ret_1 : [users=0] = get_element[node=%TopKV2_14](1)
  %Const_15 : [#users=1] = Node[type=Const] (attrs = {value: [-1]})
  %ReduceSum_16 : [#users=1] = Node[type=ReduceSum] (inputs = (x=%ret, axes=%Const_15), attrs = {keep_dims: false, noop_with_empty_axes: true})
  %Const_17 : [#users=1] = Node[type=Const] (attrs = {value: [4]})
  %TopKV2_18 : [#users=2] = Node[type=TopKV2] (inputs = (x=%ReduceSum_16, k=%Const_17), attrs = {sorted: false, dim: -1, largest: true, indices_dtype: 3})
  %ret_2 : [users=0] = get_element[node=%TopKV2_18](0)
  %ret_3 : [users=1] = get_element[node=%TopKV2_18](1)
  %Cast_19 : [#users=1] = Node[type=Cast] (inputs = (x=%ret_3), attrs = {dst_type: 9})
  %ZerosLike_20 : [#users=1] = Node[type=ZerosLike] (inputs = (x=%ReduceSum_16))
  %Shape_21 : [#users=1] = Node[type=Shape] (inputs = (x=%Cast_19), attrs = {dtype: 3})
  %Const_22 : [#users=1] = Node[type=Const] (attrs = {value: [1.000000]})
  %Cast_23 : [#users=1] = Node[type=Cast] (inputs = (x=%Const_22), attrs = {dst_type: 0})
  %Fill_24 : [#users=1] = Node[type=Fill] (inputs = (dims=%Shape_21, value=%Cast_23))
  %ScatterElements_25 : [#users=1] = Node[type=ScatterElements] (inputs = (data=%ZerosLike_20, indices=%Cast_19, updates=%Fill_24), attrs = {axis: 0, reduction: "none"})
  %Unsqueeze_26 : [#users=1] = Node[type=Unsqueeze] (inputs = (x=%ScatterElements_25), attrs = {axes: {-1}})
  %Const_27 : [#users=1] = Node[type=Const] (attrs = {value: [256 256]})
  %BroadcastTo_28 : [#users=1] = Node[type=BroadcastTo] (inputs = (x=%Unsqueeze_26, shape=%Const_27))
  %Identity_29 : [#users=1] = Node[type=Identity] (inputs = (x=%BroadcastTo_28))
  %Const_30 : [#users=1] = Node[type=Const] (attrs = {value: [256 256]})
  %Reshape_31 : [#users=1] = Node[type=Reshape] (inputs = (x=%Identity_29, shape=%Const_30), attrs = {axis: 0, num_axes: -1})
  %Cast_32 : [#users=1] = Node[type=Cast] (inputs = (x=%Reshape_31), attrs = {dst_type: 12})
  %LogicalNot_33 : [#users=1] = Node[type=LogicalNot] (inputs = (x=%Cast_32))
  %Const_34 : [#users=1] = Node[type=Const] (attrs = {value: [0.000000]})
  %MaskedFill_35 : [#users=1] = Node[type=MaskedFill] (inputs = (x=%Add_12, mask=%LogicalNot_33, value=%Const_34))
  %Const_36 : [#users=1] = Node[type=Const] (attrs = {value: [4]})
  %TopKV2_37 : [#users=2] = Node[type=TopKV2] (inputs = (x=%MaskedFill_35, k=%Const_36), attrs = {sorted: false, dim: -1, largest: true, indices_dtype: 3})
  %ret_4 : [users=0] = get_element[node=%TopKV2_37](0)
  %ret_5 : [users=1] = get_element[node=%TopKV2_37](1)
  %Cast_38 : [#users=1] = Node[type=Cast] (inputs = (x=%ret_5), attrs = {dst_type: 9})
  %GatherElements_39 : [#users=1] = Node[type=GatherElements] (inputs = (x=%Sigmoid_7, index=%Cast_38), attrs = {dim: 1})
  %Const_40 : [#users=1] = Node[type=Const] (attrs = {value: [0.000001]})
  %RealDiv_41 : [#users=1] = Node[type=RealDiv] (inputs = (x1=%GatherElements_39, x2=%Const_40))
  %Const_42 : [#users=1] = Node[type=Const] (attrs = {value: [2.500000]})
  %Mul_43 : [#users=1] = Node[type=Mul] (inputs = (x1=%RealDiv_41, x2=%Const_42))
  %Cast_44 : [#users=1] = Node[type=Cast] (inputs = (x=%Mul_43), attrs = {dst_type: 0})

  return (output_0=%Cast_38, output_1=%Cast_44)

Example 2: Below is a file containing subgraphs:

graph("TransformerBlockSubgraph"):
  %input_0 : [#users=1] = Node[type=Data] (attrs = {index: 0})
  %pred_1 : [#users=1] = Node[type=Data] (attrs = {index: 1})
  %If_0 : [#users=1] = Node[type=If] (inputs = (cond=%pred_1, input_0=%input_0), attrs = {then_branch: %If_then, else_branch: %If_else})
  %Const_1 : [#users=1] = Node[type=Const] (attrs = {value: [0]})
  %Const_2 : [#users=1] = Node[type=Const] (attrs = {value: [4]})
  %Const_3 : [#users=1] = Node[type=Const] (attrs = {value: [1]})
  %For_6 : [#users=2] = Node[type=For] (inputs = (start=%Const_1, limit=%Const_2, delta=%Const_3, input_1=%If_0), attrs = {body: %For_body})
  %ret : [users=1] = get_element[node=%For_6](0)
  %ret_1 : [users=1] = get_element[node=%For_6](1)

  return (output_0=%ret, output_1=%ret_1)

graph("If_then"):
  %input_0 : [#users=1] = Node[type=Data] (attrs = {index: 0})
  %Const_0 : [#users=1] = Node[type=Const] (attrs = {value: [0.900000]})
  %Mul_1 : [#users=1] = Node[type=Mul] (inputs = (x1=%input_0, x2=%Const_0))
  return (%Mul_1)

graph("If_else"):
  %input_0 : [#users=1] = Node[type=Data] (attrs = {index: 0})
  %Identity_0 : [#users=1] = Node[type=Identity] (inputs = (x=%input_0))
  return (%Identity_0)

graph("For_body"):
  %iter : [#users=1] = Node[type=Data] (attrs = {index: 0})
  %hidden : [#users=1] = Node[type=Data] (attrs = {index: 1})
  %Const_0 : [#users=1] = Node[type=Const] (attrs = {value: [1]})
  %Add_0 : [#users=1] = Node[type=Add] (inputs = (x1=%iter, x2=%Const_0))
  %Const_1 : [#users=1] = Node[type=Const] (attrs = {value: [0.500000]})
  %Mul_1 : [#users=1] = Node[type=Mul] (inputs = (x1=%hidden, x2=%Const_1))
  %Add_2 : [#users=1] = Node[type=Add] (inputs = (x1=%hidden, x2=%Mul_1))

  return (output_0=%Add_0, output_1=%Add_2)

Field explanations:

Graph name: graph("<graph_name>").
Node instance: %<node_instance_name> : [#users=<out_degree>] = Node[type=<node_type>](inputs = (<input_name1>=%<input_instance1>, ...), attrs = {<attribute_name1>: <attribute_value1>, ...}).
- %<node_instance_name>: Node instance name.
- [#users=<out_degree>]: Number of node outputs.
- Node[type=<node_type>]: Operator type corresponding to the node, for example MatMul node displays as Node[type=MatMul].
- inputs = (<input_name1>=%<input_instance1>, ...): Node inputs are displayed in "parameter_name=instance_name" format; when parameter name parsing fails, it falls back to _input_N sequence. This field is omitted when the node has no inputs.
  - Dynamic inputs: Dynamic input parameters use input_name_#cnt numbering format (input_name_0, input_name_1, ...).
- attrs = {<attribute_name1>: <attribute_value1>, ...}: Node attribute collection, containing subgraph attribute items and regular attribute items. This field is omitted when attributes are empty.
Multi-output node output reference: %ret/ret_#cnt : [#users=<consumer_count>] = get_element[node=%<node_instance_name>](<output_index>).
- %ret/ret_#cnt: Each output of a multi-output node uses the unified ret naming ret, ret_1, ret_2, ....
- [#users=<consumer_count>]: Number of consumers for this output.
- get_element[node=%<node_instance_name>](<output_index>): Uses get_element to represent "extract the <output_index>-th output from a multi-output node".
Graph output: return (<output_list>) Represents graph output, corresponding to NetOutput node inputs.
- Single output: return (%<output_instance>).
- Multiple outputs: return (output_0=%<output_instance0>, output_1=%<output_instance1>, ...). Outputs use output_#cnt numbering format (output_0, output_1, ...).
Subgraph representation: Nodes containing subgraphs are represented according to the following rules:
- Subgraph declaration: Declared in parent node's attrs: attrs = {<subgraph_attribute_name>: %<subgraph_instance_name>, ... }; when subgraph attribute name parsing fails, it falls back to _graph_N sequence.
- Input correspondence: Parent node input input_#cnt (or args_#cnt) corresponds to the Data node with index attribute value of cnt in the subgraph. For example: Parent node's input_0 → Subgraph's Data(attrs = {index: 0}).
- Output correspondence:
  - Single output: Subgraph return value directly serves as parent node output.
  - Multiple outputs: Subgraph output output_#cnt corresponds to parent node's cnt-th output. Parent node extracts corresponding output via get_element[node=%<parent_node>](cnt).
- Subgraph display position: Subgraph content is displayed separately after the main parent graph output ends; each subgraph starts with graph("<subgraph_instance_name>"):.

Appendix

Const Constant Display Rules

To balance readability and information completeness, the value attribute of Const nodes is displayed with the following abbreviation rules:

No value attribute: Not displayed
Empty tensor: Indicates no data content, displayed as value: <empty>
Single value: Represents scalar [] or single element [1], displays the actual value directly, e.g. value: [1.5]
Multiple values: When values exceed 6, use abbreviated form keeping first 3 and last 3 values, e.g. value: [1 2 3 ... 98 99 100]
Unsupported data type: Displayed as value: <not_supported>