blob: 29ffee867b1dc41a7c986348f3d240a9855b8cf0 [file] [log] [blame]
Sang-Hoon Park0d008f72020-03-13 14:56:05 +00001/*
2 * Copyright (c) 2020 ARM Limited.
3 *
4 * SPDX-License-Identifier: MIT
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining a copy
7 * of this software and associated documentation files (the "Software"), to
8 * deal in the Software without restriction, including without limitation the
9 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10 * sell copies of the Software, and to permit persons to whom the Software is
11 * furnished to do so, subject to the following conditions:
12 *
13 * The above copyright notice and this permission notice shall be included in all
14 * copies or substantial portions of the Software.
15 *
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22 * SOFTWARE.
23 */
24#include "arm_compute/core/NEON/kernels/NEQLSTMLayerNormalizationKernel.h"
25
26#include "arm_compute/core/CPP/Validate.h"
27#include "arm_compute/core/Helpers.h"
28#include "arm_compute/core/NEON/NEFixedPoint.h"
29#include "arm_compute/core/NEON/NEMath.h"
30#include "arm_compute/core/NEON/NESymm.h"
31#include "arm_compute/core/NEON/kernels/detail/NEActivationFunctionDetail.h"
32#include "arm_compute/core/TensorInfo.h"
33#include "arm_compute/core/Utils.h"
34#include "arm_compute/core/Validate.h"
35#include "arm_compute/core/Window.h"
36#include "arm_compute/core/utils/quantization/AsymmHelpers.h"
37
38#include <map>
39
40namespace arm_compute
41{
42namespace
43{
44inline std::pair<int64_t, int64_t> compute_mean_variance(int64_t sum, int64_t sum_sq, uint32_t num_input)
45{
46 const auto temp = static_cast<int64_t>(0x100000) / num_input;
47 const auto mean = sum * 1024 / static_cast<int64_t>(num_input);
48 const int64_t variance = ((sum_sq * temp) - (mean * mean)) / 0x100000;
49
50 return std::make_pair(mean, variance);
51}
52
53inline int64x2x2_t mul_add(const int32x4_t &a, const int32x4_t &b, const int32x4_t &bias)
54{
55 using namespace wrapper;
56 const int64x2_t a_low = vmovl(vgetlow(a));
57 const int64x2_t a_high = vmovl(vgethigh(a));
58 const int64x2_t b_low = vmovl(vgetlow(b));
59 const int64x2_t b_high = vmovl(vgethigh(b));
60
61 const int64_t a_0 = vgetlane(a_low, 0);
62 const int64_t a_1 = vgetlane(a_low, 1);
63 const int64_t a_2 = vgetlane(a_high, 0);
64 const int64_t a_3 = vgetlane(a_high, 1);
65
66 const int64_t b_0 = vgetlane(b_low, 0);
67 const int64_t b_1 = vgetlane(b_low, 1);
68 const int64_t b_2 = vgetlane(b_high, 0);
69 const int64_t b_3 = vgetlane(b_high, 1);
70
71 int64x2x2_t result;
72 const int64x2_t result_0{ a_0 * b_0, a_1 * b_1 };
73 const int64x2_t result_1{ a_2 * b_2, a_3 * b_3 };
74 result.val[0] = vadd(vmovl(vgetlow(bias)), result_0);
75 result.val[1] = vadd(vmovl(vgethigh(bias)), result_1);
76
77 return result;
78}
79} // namespace
80
81void NEQLSTMLayerNormalizationKernel::configure(const ITensor *input, ITensor *output, const ITensor *weight, const ITensor *bias)
82{
Sang-Hoon Park9230e272020-04-18 00:46:34 +010083 ARM_COMPUTE_ERROR_ON_NULLPTR(input, weight, bias, output);
84 ARM_COMPUTE_ERROR_ON(input == output);
85 ARM_COMPUTE_ERROR_THROW_ON(validate(input->info(), output->info(), weight->info(), bias->info()));
Sang-Hoon Park0d008f72020-03-13 14:56:05 +000086
87 static const std::map<DataType, ComputeFuncType> fn_map =
88 {
89 { DataType::QSYMM16, std::mem_fn(&NEQLSTMLayerNormalizationKernel::compute_qsymm16) },
90 };
91
92 _input = input;
93 _output = output;
94 _weight = weight;
95 _bias = bias;
96 _fn = fn_map.at(_input->info()->data_type());
97
98 auto_init_if_empty(*_output->info(), *_input->info());
Sang-Hoon Park9230e272020-04-18 00:46:34 +010099 _output->info()->set_quantization_info(compute_output_qinfo());
Sang-Hoon Park0d008f72020-03-13 14:56:05 +0000100
101 const UniformQuantizationInfo wq_info = _weight->info()->quantization_info().uniform();
102 const Status s = quantization::calculate_quantized_multiplier(wq_info.scale, &_output_multiplier, &_output_shift);
103 _output_shift *= -1;
104
105 if(!bool(s))
106 {
107 _output_multiplier = 0;
108 _output_shift = 0;
109 }
110
111 Window win = configure_window(output);
112 INEKernel::configure(win);
113}
114
115Window NEQLSTMLayerNormalizationKernel::configure_window(ITensor *target)
116{
117 Window window = calculate_max_window(*target->info(), Steps());
118 Coordinates coord;
119 coord.set_num_dimensions(target->info()->num_dimensions());
120 target->info()->set_valid_region(ValidRegion(coord, target->info()->tensor_shape()));
121
122 _window_start_x = static_cast<int32_t>(window.x().start());
123 _window_end_x = static_cast<int32_t>(window.x().end());
124 _window_step_x = static_cast<int32_t>(vector_size_byte) / _output->info()->element_size();
125
126 // input and output windows will iterator over y-axis, while execute_window will handler x-axis.
127 _inout_window = window;
128 _inout_window.set(Window::DimX, Window::Dimension(0, 1, 1));
129
130 // weight and bias cannot iterator along y-axis since they are 1D.
131 _weight_window = _inout_window;
132 _weight_window.set(Window::DimY, Window::Dimension(0, 1, 1));
133
134 return window;
135}
136
137Status NEQLSTMLayerNormalizationKernel::validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *weight, const ITensorInfo *bias)
138{
139 ARM_COMPUTE_UNUSED(output, bias, weight, input);
140
141 ARM_COMPUTE_ERROR_ON_NULLPTR(input, weight, bias, output);
142
143 ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::QSYMM16);
144 ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(weight, 1, DataType::QSYMM16);
145 ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(bias, 1, DataType::S32);
146
147 ARM_COMPUTE_RETURN_ERROR_ON(input->num_dimensions() > max_input_dimension);
148 ARM_COMPUTE_RETURN_ERROR_ON(weight->num_dimensions() > max_weight_dimension);
149 ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > max_bias_dimension);
150
151 ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape().x() != weight->tensor_shape().x());
152 ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(weight, bias);
153
154 if(output->total_size() != 0)
155 {
156 ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
157 ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(input, output);
158 }
159
160 return Status{};
161}
162
163void NEQLSTMLayerNormalizationKernel::run(const Window &window, const ThreadInfo &info)
164{
165 ARM_COMPUTE_UNUSED(window, info);
166 ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
167 ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window);
168 ARM_COMPUTE_ERROR_ON_MSG(!_fn, "internal function is not defined for computation");
169
170 _fn(*this);
171}
172
Sang-Hoon Park9230e272020-04-18 00:46:34 +0100173inline QuantizationInfo NEQLSTMLayerNormalizationKernel::compute_output_qinfo()
174{
Sheri Zhang3a353982020-04-21 13:10:24 +0100175 return QuantizationInfo(1.f / 4096);
Sang-Hoon Park9230e272020-04-18 00:46:34 +0100176}
177
Sang-Hoon Park0d008f72020-03-13 14:56:05 +0000178inline std::pair<int64_t, int64_t> NEQLSTMLayerNormalizationKernel::sum_qsymm16(const int16_t *input_ptr)
179{
180 ARM_COMPUTE_ERROR_ON(!input_ptr);
181
182 using AccType = int64_t;
183 using InputDataType = int16_t;
184
185 AccType sum{ 0 };
186 AccType sum_sq{ 0 };
187
188 int32_t x = _window_start_x;
189 for(; x <= _window_end_x && _window_step_x <= (_window_end_x - x); x += _window_step_x)
190 {
191 using namespace wrapper;
192 const int16x8_t val = vloadq(input_ptr + x);
193 const int32x4_t val_low = vmovl(vgetlow(val));
194 const int32x4_t val_high = vmovl(vgethigh(val));
195
196#if defined(__aarch64__)
197 sum += static_cast<AccType>(vaddv(val_low));
198 sum += static_cast<AccType>(vaddv(val_high));
199
200 sum_sq += static_cast<AccType>(vaddv(vmul(val_low, val_low)));
201 sum_sq += static_cast<AccType>(vaddv(vmul(val_high, val_high)));
202#else // __aarch64__
203 // only AArch64 supports vaddv
204 const int64x2_t pair_sum_low = vpaddl(val_low);
205 const int64x2_t pair_sum_high = vpaddl(val_high);
206 const int64x2_t pair_sum = vadd(pair_sum_low, pair_sum_high);
207 sum += vgetlane(pair_sum, 0) + vgetlane(pair_sum, 1);
208
209 const int32x4_t square_low = vmul(val_low, val_low);
210 const int32x4_t square_high = vmul(val_high, val_high);
211 const int64x2_t pair_sum_sq_low = vpaddl(square_low);
212 const int64x2_t pair_sum_sq_high = vpaddl(square_high);
213 const int64x2_t pair_sum_sq = vadd(pair_sum_sq_low, pair_sum_sq_high);
214 sum_sq += vgetlane(pair_sum_sq, 0) + vgetlane(pair_sum_sq, 1);
215#endif // __aarch64__
216 }
217
218 for(; x < _window_end_x; ++x)
219 {
220 const InputDataType val = input_ptr[x];
221 sum += static_cast<AccType>(val);
222 sum_sq += static_cast<AccType>(val * val);
223 }
224
225 return std::make_pair(sum, sum_sq);
226}
227
228inline void NEQLSTMLayerNormalizationKernel::normalize_qasymm16(const int16_t *input_ptr,
229 int16_t *output_ptr,
230 const int16_t *weight_ptr,
231 const int32_t *bias_ptr,
232 int32_t mean, int32_t inv_std_mul, int32_t inv_std_shift)
233{
234 using OutputDataType = int16_t;
235
236 using namespace wrapper;
237 const int32x4_t mean_vec = vdup_n(mean, wrapper::traits::vector_128_tag{});
238
239 int32_t x = _window_start_x;
240 for(; x <= _window_end_x && _window_step_x <= (_window_end_x - x); x += _window_step_x)
241 {
242 const int16x8_t val = vloadq(input_ptr + x);
243 int32x4x2_t shifted;
244 shifted.val[0] = vsub(vshlq_n_s32(vmovl(vgetlow(val)), 10), mean_vec);
245 shifted.val[1] = vsub(vshlq_n_s32(vmovl(vgethigh(val)), 10), mean_vec);
246
247 int32x4x2_t rescaled = multiply_by_quantized_multiplier_2row(shifted, inv_std_mul, inv_std_shift);
248
249 const int16x8_t weight_val = vloadq(weight_ptr + x);
250 const int32x4_t weight_low = vmovl(vgetlow(weight_val));
251 const int32x4_t weight_high = vmovl(vgethigh(weight_val));
252
253 const int32x4_t bias_low = vloadq(bias_ptr + x);
254 const int32x4_t bias_high = vloadq(bias_ptr + 4 + x);
255
256 int64x2x2_t result_0 = mul_add(rescaled.val[0], weight_low, bias_low);
257 int64x2x2_t result_1 = mul_add(rescaled.val[1], weight_high, bias_high);
258
259 int32x4x2_t combined;
260 combined.val[0] = vcombine(vmovn(vrshrq_n_s64(result_0.val[0], 10)), vmovn(vrshrq_n_s64(result_0.val[1], 10)));
261 combined.val[1] = vcombine(vmovn(vrshrq_n_s64(result_1.val[0], 10)), vmovn(vrshrq_n_s64(result_1.val[1], 10)));
262
263 int32x4x2_t out_val = multiply_by_quantized_multiplier_2row(combined, _output_multiplier, _output_shift + 12);
264
265 vstore(output_ptr + x, vqmovn(out_val.val[0]));
266 vstore(output_ptr + x + 4, vqmovn(out_val.val[1]));
267 }
268
269 for(; x < _window_end_x; ++x)
270 {
271 const auto val = static_cast<int32_t>(input_ptr[x]);
272 const int32_t shifted = (val << 10) - mean;
273 const int32_t rescaled = quantization::multiply_by_quantized_multiplier(shifted, inv_std_mul, inv_std_shift);
274 const int64_t weighted = rescaled * weight_ptr[x] + bias_ptr[x];
275 const auto reverse_shifted = static_cast<int32_t>((weighted + 512) >> 10);
276 int32_t out_val = quantization::multiply_by_quantized_multiplier(reverse_shifted, _output_multiplier, _output_shift + 12);
277 out_val = utility::clamp<decltype(out_val), OutputDataType>(out_val, std::numeric_limits<OutputDataType>::min());
278 output_ptr[x] = static_cast<OutputDataType>(out_val);
279 }
280}
281
282void NEQLSTMLayerNormalizationKernel::compute_qsymm16()
283{
284 using InputDataType = int16_t;
285 using OutputDataType = int16_t;
286 using BiasDataType = int32_t;
287 using AccType = int64_t;
288
289 Iterator input_iterator{ _input, _inout_window };
290 Iterator output_iterator{ _output, _inout_window };
291 Iterator weight_iterator{ _weight, _weight_window };
292 Iterator bias_iterator{ _bias, _weight_window };
293
294 const auto weight_ptr = reinterpret_cast<const InputDataType *>(weight_iterator.ptr());
295 const auto bias_ptr = reinterpret_cast<const BiasDataType *>(bias_iterator.ptr());
296
297 const uint32_t column_size = _input->info()->tensor_shape()[0];
298
299 execute_window_loop(_inout_window, [ &, this](const Coordinates &)
300 {
301 const auto in_ptr = reinterpret_cast<const InputDataType *>(input_iterator.ptr());
302 auto out_ptr = reinterpret_cast<OutputDataType *>(output_iterator.ptr());
303
304 AccType sum{ 0 };
305 AccType sum_sq{ 0 };
306 std::tie(sum, sum_sq) = sum_qsymm16(in_ptr);
307
308 AccType mean{ 0 };
309 AccType variance{ 0 };
310 std::tie(mean, variance) = compute_mean_variance(sum, sum_sq, column_size);
311
312 int32_t stddev_invsqrt_mul{};
313 int32_t stddev_invsqrt_shift{};
314 quantization::get_invsqrt_quantized_multiplier_exp(static_cast<int32_t>(variance), -1, stddev_invsqrt_mul, stddev_invsqrt_shift);
315
316 normalize_qasymm16(in_ptr, out_ptr, weight_ptr, bias_ptr, mean, stddev_invsqrt_mul, stddev_invsqrt_shift);
317 },
318 input_iterator, output_iterator);
319}
320} // namespace arm_compute