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
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
|
/* 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 <cmath>
#include "tensorflow/compiler/tf2xla/lib/random.h"
#include "tensorflow/compiler/tf2xla/shape_util.h"
#include "tensorflow/compiler/tf2xla/xla_helpers.h"
#include "tensorflow/compiler/tf2xla/xla_op_kernel.h"
#include "tensorflow/compiler/tf2xla/xla_op_registry.h"
#include "tensorflow/compiler/xla/client/lib/arithmetic.h"
#include "tensorflow/compiler/xla/client/xla_client/xla_builder.h"
#include "tensorflow/core/framework/op_kernel.h"
#include "tensorflow/core/framework/tensor.h"
#include "tensorflow/core/framework/tensor_shape.h"
#include "tensorflow/core/lib/core/casts.h"
#include "tensorflow/core/lib/math/math_util.h"
namespace tensorflow {
namespace {
// Rotates a 32-bit integer 'v' left by 'distance' bits.
xla::XlaOp RotateLeftS32(xla::XlaBuilder* builder, const xla::XlaOp& v,
int distance) {
return xla::Or(
xla::ShiftLeft(v, xla::ConstantR0<int>(builder, distance)),
xla::ShiftRightLogical(v, xla::ConstantR0<int>(builder, 32 - distance)));
}
using ThreeFry2x32State = std::array<xla::XlaOp, 2>;
// Implements the ThreeFry counter-based PRNG algorithm.
// Salmon et al. SC 2011. Parallel random numbers: as easy as 1, 2, 3.
// http://www.thesalmons.org/john/random123/papers/random123sc11.pdf
ThreeFry2x32State ThreeFry2x32(xla::XlaBuilder* builder,
ThreeFry2x32State input, ThreeFry2x32State key) {
// Rotation distances specified by the Threefry2x32 algorithm.
constexpr std::array<int, 8> rotations = {13, 15, 26, 6, 17, 29, 16, 24};
ThreeFry2x32State x;
std::array<xla::XlaOp, 3> ks;
// 0x1BD11BDA is a parity constant specified by the ThreeFry2x32 algorithm.
ks[2] = xla::ConstantR0<int32>(builder, 0x1BD11BDA);
for (int i = 0; i < 2; ++i) {
ks[i] = key[i];
x[i] = input[i];
ks[2] = xla::Xor(ks[2], key[i]);
}
x[0] = xla::Add(x[0], ks[0]);
x[1] = xla::Add(x[1], ks[1]);
// Performs a single round of the Threefry2x32 algorithm, with a rotation
// amount 'rotation'.
auto round = [builder](ThreeFry2x32State v, int rotation) {
v[0] = xla::Add(v[0], v[1]);
v[1] = RotateLeftS32(builder, v[1], rotation);
v[1] = xla::Xor(v[0], v[1]);
return v;
};
// There are no known statistical flaws with 13 rounds of Threefry2x32.
// We are conservative and use 20 rounds.
x = round(x, rotations[0]);
x = round(x, rotations[1]);
x = round(x, rotations[2]);
x = round(x, rotations[3]);
x[0] = xla::Add(x[0], ks[1]);
x[1] = xla::Add(xla::Add(x[1], ks[2]), xla::ConstantR0<int32>(builder, 1));
x = round(x, rotations[4]);
x = round(x, rotations[5]);
x = round(x, rotations[6]);
x = round(x, rotations[7]);
x[0] = xla::Add(x[0], ks[2]);
x[1] = xla::Add(xla::Add(x[1], ks[0]), xla::ConstantR0<int32>(builder, 2));
x = round(x, rotations[0]);
x = round(x, rotations[1]);
x = round(x, rotations[2]);
x = round(x, rotations[3]);
x[0] = xla::Add(x[0], ks[0]);
x[1] = xla::Add(xla::Add(x[1], ks[1]), xla::ConstantR0<int32>(builder, 3));
x = round(x, rotations[4]);
x = round(x, rotations[5]);
x = round(x, rotations[6]);
x = round(x, rotations[7]);
x[0] = xla::Add(x[0], ks[1]);
x[1] = xla::Add(xla::Add(x[1], ks[2]), xla::ConstantR0<int32>(builder, 4));
x = round(x, rotations[0]);
x = round(x, rotations[1]);
x = round(x, rotations[2]);
x = round(x, rotations[3]);
x[0] = xla::Add(x[0], ks[2]);
x[1] = xla::Add(xla::Add(x[1], ks[0]), xla::ConstantR0<int32>(builder, 5));
return x;
}
// Returns a tensor of 'shape' random values uniformly distributed in the range
// [minval, maxval)
xla::XlaOp RandomUniform(xla::XlaBuilder* builder, const xla::XlaOp& seed,
const TensorShape& shape, double minval,
double maxval) {
// Split the seed into two 32-bit scalars to form a key.
auto seed0 = xla::Reshape(xla::Slice(seed, {0}, {1}, {1}), {});
auto seed1 = xla::Reshape(xla::Slice(seed, {1}, {2}, {1}), {});
ThreeFry2x32State key = {seed0, seed1};
const int64 size = shape.num_elements();
const int64 half_size = MathUtil::CeilOfRatio<int64>(size, 2);
const bool size_is_odd = (half_size * 2 != size);
// Fill the generator inputs with unique counter values.
ThreeFry2x32State inputs;
TF_CHECK_OK(XlaHelpers::Iota(builder, DT_INT32, half_size, &inputs[0]));
inputs[1] = xla::Add(inputs[0], xla::ConstantR0<int32>(builder, half_size));
ThreeFry2x32State outputs = ThreeFry2x32(builder, inputs, key);
if (size_is_odd) {
outputs[1] = xla::Slice(outputs[1], {0}, {half_size - 1}, {1});
}
auto bits =
xla::Reshape(xla::ConcatInDim(builder, outputs, 0), shape.dim_sizes());
// Form 22 random mantissa bits, with a leading 1 bit. The leading 1 bit
// forces the random bits into the mantissa.
constexpr int kFloatBits = 32;
constexpr int kMantissaBits = 23;
bits = xla::Or(
xla::ShiftRightLogical(
bits, xla::ConstantR0<int32>(builder, kFloatBits - kMantissaBits)),
xla::ConstantR0<int32>(builder, bit_cast<int32>(1.0f)));
auto floats = xla::BitcastConvertType(bits, xla::F32);
// We have a floating point number in the range [1.0, 2.0).
// Subtract 1.0f to shift to the range [0.0, 1.0)
floats = xla::Sub(floats, xla::ConstantR0<float>(builder, 1.0f));
// Multiply and add to shift to the range [minval, maxval).
floats = xla::Mul(floats, xla::ConstantR0<float>(builder, maxval - minval));
floats = xla::Add(floats, xla::ConstantR0<float>(builder, minval));
return floats;
}
} // namespace
class StatelessRandomUniformOp : public XlaOpKernel {
public:
explicit StatelessRandomUniformOp(OpKernelConstruction* ctx)
: XlaOpKernel(ctx) {}
void Compile(XlaOpKernelContext* ctx) override {
xla::XlaBuilder* builder = ctx->builder();
TensorShape shape;
OP_REQUIRES_OK(ctx, ctx->ConstantInputAsShape(0, &shape));
TensorShape seed_shape = ctx->InputShape(1);
OP_REQUIRES(ctx, seed_shape.dims() == 1 && seed_shape.dim_size(0) == 2,
errors::InvalidArgument("seed must have shape [2], not ",
seed_shape.DebugString()));
xla::XlaOp seed = ctx->Input(1);
ctx->SetOutput(0, RandomUniform(builder, seed, shape, 0.0, 1.0));
}
private:
TF_DISALLOW_COPY_AND_ASSIGN(StatelessRandomUniformOp);
};
// TODO(phawkins): generalize to non-float, non-int32 seed types.
REGISTER_XLA_OP(Name("StatelessRandomUniform")
.CompileTimeConstInput("shape")
.TypeConstraint("dtype", DT_FLOAT)
.TypeConstraint("Tseed", DT_INT32),
StatelessRandomUniformOp);
class StatelessRandomNormalOp : public XlaOpKernel {
public:
explicit StatelessRandomNormalOp(OpKernelConstruction* ctx)
: XlaOpKernel(ctx) {}
void Compile(XlaOpKernelContext* ctx) override {
TensorShape shape;
OP_REQUIRES_OK(ctx, ctx->ConstantInputAsShape(0, &shape));
TensorShape seed_shape = ctx->InputShape(1);
OP_REQUIRES(ctx, seed_shape == TensorShape({2}),
errors::InvalidArgument("seed must have shape [2], not ",
seed_shape.DebugString()));
xla::XlaOp seed = ctx->Input(1);
xla::XlaBuilder* builder = ctx->builder();
auto uniform =
RandomUniform(builder, seed, shape, std::nextafter(-1.0f, 0.0f), 1.0);
// Convert uniform distribution to normal distribution by computing
// sqrt(2) * erfinv(x)
auto normal = xla::Mul(xla::ConstantR0<float>(builder, std::sqrt(2.0)),
ErfInv(uniform));
ctx->SetOutput(0, normal);
}
private:
TF_DISALLOW_COPY_AND_ASSIGN(StatelessRandomNormalOp);
};
// TODO(phawkins): generalize to non-float, non-int32 seed types.
REGISTER_XLA_OP(Name("StatelessRandomNormal")
.CompileTimeConstInput("shape")
.TypeConstraint("dtype", DT_FLOAT)
.TypeConstraint("Tseed", DT_INT32),
StatelessRandomNormalOp);
class StatelessTruncatedNormalOp : public XlaOpKernel {
public:
explicit StatelessTruncatedNormalOp(OpKernelConstruction* ctx)
: XlaOpKernel(ctx) {}
void Compile(XlaOpKernelContext* ctx) override {
const DataType dtype = output_type(0);
TensorShape shape;
OP_REQUIRES_OK(ctx, ctx->ConstantInputAsShape(0, &shape));
TensorShape seed_shape = ctx->InputShape(1);
OP_REQUIRES(ctx, seed_shape == TensorShape({2}),
errors::InvalidArgument("seed must have shape [2], not ",
seed_shape.DebugString()));
xla::XlaOp seed = ctx->Input(1);
xla::XlaBuilder* b = ctx->builder();
auto uniform =
RandomUniform(b, seed, shape, std::numeric_limits<float>::min(), 1.0);
ctx->SetOutput(0, TruncatedNormal(dtype, uniform));
}
private:
TF_DISALLOW_COPY_AND_ASSIGN(StatelessTruncatedNormalOp);
};
REGISTER_XLA_OP(Name("StatelessTruncatedNormal")
.CompileTimeConstInput("shape")
.TypeConstraint("dtype", DT_FLOAT)
.TypeConstraint("Tseed", DT_INT32),
StatelessTruncatedNormalOp);
} // namespace tensorflow
|
