Files
2026-07-13 12:40:42 +08:00

282 lines
12 KiB
C++

// Copyright (c) 2022 PaddlePaddle 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 "paddle/phi/kernels/send_uv_grad_kernel.h"
#include "paddle/common/hostdevice.h"
#include "paddle/phi/backends/cpu/cpu_context.h"
#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/empty_kernel.h"
#include "paddle/phi/kernels/full_kernel.h"
#include "paddle/phi/kernels/funcs/math_function.h"
#include "paddle/phi/kernels/impl/graph_message_passing_impl.h"
#include "paddle/phi/kernels/reduce_sum_kernel.h"
namespace phi {
template <typename Context, typename T, typename IndexT>
void CalculateGrad(const Context& dev_ctx,
const T* out_grad,
const IndexT* s_index,
const IndexT* d_index,
const DDim& out_grad_dims,
const DDim& x_grad_dims,
const std::string& message_op,
int64_t index_size,
int64_t slice_size,
T* x_grad,
const DenseTensor& out_grad_tensor UNUSED,
const DenseTensor& y) {
std::vector<int64_t> reduce_idx;
bool reduce = ReduceGrad(out_grad_dims, x_grad_dims, reduce_idx);
if (message_op == "ADD") {
if (!reduce) {
#ifdef PADDLE_WITH_MKLML
#pragma omp parallel for
#endif
for (int64_t i = 0; i < index_size; i++) {
IndexT dst = d_index[i];
T* x_grad_off = x_grad + dst * slice_size;
const T* out_grad_off = out_grad + i * slice_size;
for (int64_t j = 0; j < slice_size; j++) {
if (out_grad_off[j] != 0) {
#ifdef PADDLE_WITH_MKLML
#pragma omp atomic
#endif
x_grad_off[j] += out_grad_off[j];
}
}
}
} else {
const auto& bcast_info = CalcBCastInfo(out_grad_dims, x_grad_dims);
auto out_grad_dims_1 = vectorize<int>(out_grad_dims);
std::vector<int> out_grad_dims_2(out_grad_dims_1.begin() + 1,
out_grad_dims_1.end());
out_grad_dims_2.emplace(out_grad_dims_2.begin(), x_grad_dims[0]);
DenseTensor x_grad_v2 = Empty<T, Context>(dev_ctx, out_grad_dims_2);
funcs::SetConstant<Context, T>()(dev_ctx, &x_grad_v2, static_cast<T>(0));
T* x_grad_v2_data = x_grad_v2.data<T>();
#ifdef PADDLE_WITH_MKLML
#pragma omp parallel for
#endif
for (int64_t i = 0; i < index_size; i++) {
IndexT dst = d_index[i];
T* x_grad_off = x_grad_v2_data + dst * bcast_info.out_len;
const T* out_grad_off = out_grad + i * bcast_info.out_len;
for (int64_t j = 0; j < bcast_info.out_len; j++) {
if (out_grad_off[j] != 0) {
#ifdef PADDLE_WITH_MKLML
#pragma omp atomic
#endif
x_grad_off[j] += out_grad_off[j];
}
}
}
DenseTensor x_grad_out = Sum<T, Context>(dev_ctx,
x_grad_v2,
IntArray(reduce_idx),
CppTypeToDataType<T>::Type(),
true);
memcpy(x_grad, x_grad_out.data<T>(), x_grad_out.numel() * sizeof(T));
}
} else if (message_op == "MUL") {
const auto& bcast = CalcBCastInfo(y.dims(), out_grad_dims);
const T* y_data = y.data<T>();
if (!reduce) {
#ifdef PADDLE_WITH_MKLML
#pragma omp parallel for
#endif
for (int64_t i = 0; i < index_size; i++) {
IndexT src = s_index[i];
IndexT dst = d_index[i];
T* x_grad_off = x_grad + dst * bcast.out_len;
const T* y_off = y_data + src * bcast.l_len;
const T* out_grad_off = out_grad + i * bcast.r_len;
for (int64_t j = 0; j < bcast.out_len; j++) {
int64_t y_add = bcast.use_bcast ? bcast.l_offset[j] : j;
int64_t o_add = bcast.use_bcast ? bcast.r_offset[j] : j;
T val = y_off[y_add] * out_grad_off[o_add];
if (val != 0) {
#ifdef PADDLE_WITH_MKLML
#pragma omp atomic
#endif
x_grad_off[j] += val;
}
}
}
} else {
auto out_grad_dims_1 = vectorize<int>(out_grad_dims);
std::vector<int> out_grad_dims_2(out_grad_dims_1.begin() + 1,
out_grad_dims_1.end());
out_grad_dims_2.emplace(out_grad_dims_2.begin(), x_grad_dims[0]);
DenseTensor x_grad_v2 = Empty<T, Context>(dev_ctx, out_grad_dims_2);
funcs::SetConstant<Context, T>()(dev_ctx, &x_grad_v2, static_cast<T>(0));
T* x_grad_v2_data = x_grad_v2.data<T>();
#ifdef PADDLE_WITH_MKLML
#pragma omp parallel for
#endif
for (int64_t i = 0; i < index_size; i++) {
IndexT src = s_index[i];
IndexT dst = d_index[i];
T* x_grad_off = x_grad_v2_data + dst * bcast.out_len;
const T* y_off = y_data + src * bcast.l_len;
const T* out_grad_off = out_grad + i * bcast.r_len;
for (int64_t j = 0; j < bcast.out_len; j++) {
int64_t y_add = bcast.use_bcast ? bcast.l_offset[j] : j;
int64_t o_add = bcast.use_bcast ? bcast.r_offset[j] : j;
T val = y_off[y_add] * out_grad_off[o_add];
if (val != 0) {
#ifdef PADDLE_WITH_MKLML
#pragma omp atomic
#endif
x_grad_off[j] += val;
}
}
}
DenseTensor x_grad_out = Sum<T, Context>(dev_ctx,
x_grad_v2,
IntArray(reduce_idx),
CppTypeToDataType<T>::Type(),
true);
memcpy(x_grad, x_grad_out.data<T>(), x_grad_out.numel() * sizeof(T));
}
}
}
template <typename Context, typename T, typename IndexT>
void GraphSendUVGradOpKernelLaunchHelper(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
const DenseTensor& out_grad,
const DenseTensor& src_index,
const DenseTensor& dst_index,
const std::string& message_op,
DenseTensor* x_grad,
DenseTensor* y_grad) {
const int64_t& index_size = dst_index.dims()[0];
PADDLE_ENFORCE_GT(
index_size,
0,
errors::InvalidArgument("The first dimension of src_index or dst_index "
"should be greater than 0, but received %d.",
index_size));
dev_ctx.template Alloc<T>(x_grad);
T* x_grad_data = x_grad->data<T>();
dev_ctx.template Alloc<T>(y_grad);
T* y_grad_data = y_grad->data<T>();
const auto& x_grad_dims = x_grad->dims();
const auto& y_grad_dims = y_grad->dims();
int64_t memset_size_x = 1, memset_size_y = 1;
int64_t slice_size_x = 1, slice_size_y = 1;
for (int i = 0; i < x_grad_dims.size(); i++) {
memset_size_x *= x_grad_dims[i];
if (i > 0) slice_size_x *= x_grad_dims[i];
}
for (int i = 0; i < y_grad_dims.size(); i++) {
memset_size_y *= y_grad_dims[i];
if (i > 0) slice_size_y *= y_grad_dims[i];
}
const size_t& memset_bytes_x = memset_size_x * sizeof(T);
const size_t& memset_bytes_y = memset_size_y * sizeof(T);
memset(x_grad_data, 0, memset_bytes_x);
memset(y_grad_data, 0, memset_bytes_y);
const T* out_grad_data = out_grad.data<T>();
const IndexT* s_index = src_index.data<IndexT>();
const IndexT* d_index = dst_index.data<IndexT>();
const auto& out_grad_dims = out_grad.dims();
// Calculate X Grad.
CalculateGrad<Context, T, IndexT>(dev_ctx,
out_grad_data,
d_index,
s_index,
out_grad_dims,
x_grad_dims,
message_op,
index_size,
slice_size_x,
x_grad_data,
out_grad,
y);
// Calculate Y Grad.
CalculateGrad<Context, T, IndexT>(dev_ctx,
out_grad_data,
s_index,
d_index,
out_grad_dims,
y_grad_dims,
message_op,
index_size,
slice_size_y,
y_grad_data,
out_grad,
x);
}
template <typename T, typename Context>
void SendUVGradKernel(const Context& dev_ctx,
const DenseTensor& x,
const DenseTensor& y,
const DenseTensor& src_index,
const DenseTensor& dst_index,
const DenseTensor& out_grad,
const std::string& message_op,
DenseTensor* x_grad,
DenseTensor* y_grad) {
auto index_type = src_index.dtype();
if (out_grad.numel() == 0 || x.numel() == 0 || y.numel() == 0 ||
src_index.numel() == 0 || dst_index.numel() == 0) {
Full<T, Context>(dev_ctx, x_grad->dims(), 0, x_grad);
Full<T, Context>(dev_ctx, y_grad->dims(), 0, y_grad);
return;
}
if (index_type == DataType::INT32) {
GraphSendUVGradOpKernelLaunchHelper<Context, T, int32_t>(dev_ctx,
x,
y,
out_grad,
src_index,
dst_index,
message_op,
x_grad,
y_grad);
} else if (index_type == DataType::INT64) {
GraphSendUVGradOpKernelLaunchHelper<Context, T, int64_t>(dev_ctx,
x,
y,
out_grad,
src_index,
dst_index,
message_op,
x_grad,
y_grad);
}
}
} // namespace phi
PD_REGISTER_KERNEL(send_uv_grad,
CPU,
ALL_LAYOUT,
phi::SendUVGradKernel,
float,
double,
int,
int64_t) {}