1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
|
/* Copyright 2017 The TensorFlow Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#include "tensorflow/compiler/xla/service/hlo_element_type_converter.h"
#include <memory>
#include <string>
#include <utility>
#include <vector>
#include "tensorflow/compiler/xla/layout_util.h"
#include "tensorflow/compiler/xla/literal.h"
#include "tensorflow/compiler/xla/service/dfs_hlo_visitor_with_default.h"
#include "tensorflow/compiler/xla/service/hlo_computation.h"
#include "tensorflow/compiler/xla/service/hlo_evaluator.h"
#include "tensorflow/compiler/xla/service/hlo_instruction.h"
#include "tensorflow/compiler/xla/service/hlo_opcode.h"
#include "tensorflow/compiler/xla/service/hlo_query.h"
#include "tensorflow/compiler/xla/shape_util.h"
#include "tensorflow/compiler/xla/types.h"
#include "tensorflow/core/lib/core/errors.h"
namespace xla {
namespace {
HloInstruction* ToElementType(HloInstruction* hlo, PrimitiveType type) {
if (hlo->shape().element_type() != type) {
Shape shape = ShapeUtil::ChangeElementType(hlo->shape(), type);
hlo = hlo->parent()->AddInstruction(
HloInstruction::CreateConvert(shape, hlo));
}
CHECK_EQ(hlo->shape().element_type(), type);
return hlo;
}
bool HasOperandType(HloInstruction* hlo, PrimitiveType type) {
for (HloInstruction* operand : hlo->operands()) {
if (operand->shape().element_type() == type) {
return true;
}
}
return false;
}
// Finds out the Tuple Shape of the new instruction after converting the element
// type of the operands of the original instruction from `from_type` to
// `to_type`.
//
// This routine assumes the resulting `shape` of the original instruction is a
// non-nested tuple. This assumption is currently safe as only kTuple, kInfeed,
// kOutfeed, kCall, kCustomCall and kBatchNorm* HLO instructions can produce
// results with tuple shapes, and this routine is only called to convert the
// result shapes of kBatchNorm* HLO instructions, which are non-nested tuples.
Shape GetConvertedTupleShape(const Shape& shape, PrimitiveType from_type,
PrimitiveType to_type) {
std::vector<Shape> new_tuple_subshapes;
for (int64 i = 0; i < ShapeUtil::TupleElementCount(shape); ++i) {
Shape subshape = ShapeUtil::GetTupleElementShape(shape, i);
CHECK(!ShapeUtil::IsTuple(subshape));
if (subshape.element_type() == from_type) {
subshape = ShapeUtil::ChangeElementType(subshape, to_type);
}
new_tuple_subshapes.push_back(subshape);
}
return ShapeUtil::MakeTupleShape(new_tuple_subshapes);
}
// Converts the elements of the result of `hlo` to produce a new tuple with
// shape `to_shape`.
//
// This routine assumes `hlo` is an instruction that produces a non-nested Tuple
// as a result.
HloInstruction* ConvertTupleElements(HloInstruction* hlo,
const Shape& to_shape) {
const Shape& shape = hlo->shape();
HloComputation* computation = hlo->parent();
std::vector<HloInstruction*> tuple_elements;
for (int64 i = 0; i < ShapeUtil::TupleElementCount(shape); ++i) {
const Shape& ele_shape = ShapeUtil::GetTupleElementShape(shape, i);
HloInstruction* element = computation->AddInstruction(
HloInstruction::CreateGetTupleElement(ele_shape, hlo, i));
const Shape& to_ele_shape = ShapeUtil::GetTupleElementShape(to_shape, i);
CHECK(!ShapeUtil::IsTuple(ele_shape));
if (ele_shape.element_type() != to_ele_shape.element_type()) {
element = computation->AddInstruction(
HloInstruction::CreateConvert(to_ele_shape, element));
}
tuple_elements.push_back(element);
}
return computation->AddInstruction(
HloInstruction::CreateTuple(tuple_elements));
}
} // namespace
HloElementTypeConverter::HloElementTypeConverter(
PrimitiveType eliminate_type, PrimitiveType replace_with_type)
: eliminate_type_(eliminate_type), replace_with_type_(replace_with_type) {}
// This routine converts the arithmetic operations in the given module that use
// eliminate_type_ to operations that use replace_with_type_.
StatusOr<bool> HloElementTypeConverter::Run(HloModule* module) {
XLA_VLOG_LINES(
3, "HloElementTypeConverter::Run(), before:\n" + module->ToString());
if (eliminate_type_ == replace_with_type_) {
return false;
}
HloCloneContext context(module);
bool changed = false;
for (auto* computation : module->computations()) {
for (auto* hlo : computation->MakeInstructionPostOrder()) {
const auto opcode = hlo->opcode();
// These are ops where it does not make sense to convert them.
if (opcode == HloOpcode::kParameter || opcode == HloOpcode::kConstant ||
opcode == HloOpcode::kTuple || opcode == HloOpcode::kConvert ||
opcode == HloOpcode::kGetTupleElement ||
opcode == HloOpcode::kInfeed || opcode == HloOpcode::kOutfeed) {
continue;
}
// We cannot change a CustomCall since we have no way of adjusting the
// called binary to expect the updated type.
if (opcode == HloOpcode::kCustomCall) {
continue;
}
// These are ops with embedded computations where it suffices to convert
// the embedded computations instead of converting the ops themselves.
if (opcode == HloOpcode::kWhile || opcode == HloOpcode::kCall ||
opcode == HloOpcode::kCrossReplicaSum ||
opcode == HloOpcode::kFusion || opcode == HloOpcode::kMap ||
opcode == HloOpcode::kReduce || opcode == HloOpcode::kReduceWindow ||
opcode == HloOpcode::kScatter ||
opcode == HloOpcode::kSelectAndScatter ||
opcode == HloOpcode::kConditional) {
continue;
}
TF_RET_CHECK(hlo->called_computations().empty()) << hlo->ToString();
bool nullary = hlo->operands().empty();
bool wrong_element_type = hlo->shape().element_type() == eliminate_type_;
bool should_eliminate_type = (nullary && wrong_element_type) ||
HasOperandType(hlo, eliminate_type_);
if (!should_eliminate_type) {
// If this CHECK fires, then this was an instruction that does not take
// the elimination type as an operand but it does return it. This pass
// does not have a feature to change the output type in that case, so
// instead of silently failing to eliminate the type, it fails loudly.
TF_RET_CHECK(hlo->shape().element_type() != eliminate_type_);
continue;
}
// Handle instructions that perform arithmetic operations and contain
// operands with eliminate_type_.
//
// First, convert the operands with eliminate_type_ to operands with
// replace_with_type_.
std::vector<HloInstruction*> new_operands;
for (HloInstruction* operand : hlo->operands()) {
if (operand->shape().element_type() == eliminate_type_) {
operand = ToElementType(operand, replace_with_type_);
}
new_operands.push_back(operand);
}
// Then find out the result type of the new instruction with the same
// opcode but using the converted operands, create the new instruction,
// and convert the result of the new instruction back to match the result
// type of the original instruction.
HloInstruction* new_hlo;
if (hlo->shape().element_type() == eliminate_type_) {
Shape shape =
ShapeUtil::ChangeElementType(hlo->shape(), replace_with_type_);
new_hlo = computation->AddInstruction(
hlo->CloneWithNewOperands(shape, new_operands, &context));
TF_RETURN_IF_ERROR(new_hlo->CopyAllControlDepsFrom(hlo));
new_hlo = ToElementType(new_hlo, eliminate_type_);
} else if (ShapeUtil::IsTuple(hlo->shape())) {
Shape old_shape = hlo->shape();
Shape new_shape = GetConvertedTupleShape(hlo->shape(), eliminate_type_,
replace_with_type_);
new_hlo = computation->AddInstruction(
hlo->CloneWithNewOperands(new_shape, new_operands, &context));
TF_RETURN_IF_ERROR(new_hlo->CopyAllControlDepsFrom(hlo));
// Convert the elements of the result of `new_hlo` to produce a new
// tuple with shape `old_shape`.
new_hlo = ConvertTupleElements(new_hlo, old_shape);
} else {
new_hlo = computation->AddInstruction(
hlo->CloneWithNewOperands(hlo->shape(), new_operands, &context));
TF_RETURN_IF_ERROR(new_hlo->CopyAllControlDepsFrom(hlo));
}
TF_RETURN_IF_ERROR(hlo->ReplaceAllUsesWith(new_hlo));
TF_RETURN_IF_ERROR(hlo->DropAllControlDeps());
// NB! We want to replace and remove side effecting instructions like Rng
// as well so we can't rely HloComputation::ReplaceInstruction to reliably
// remove the replaced instruction.
TF_RETURN_IF_ERROR(computation->RemoveInstruction(hlo));
changed = true;
}
}
XLA_VLOG_LINES(
2, "HloElementTypeConverter::Run(), after:\n" + module->ToString());
return changed;
}
} // namespace xla
|
