[Phi]Add diag_v2 grad kernel (#40447)

* Add diag grad kernel

* fix unittest case

* add float16, remove const &

* delete diag_grad in op_utils.h

paddle/fluid/operators/diag_v2_op.cc

@@ -12,8 +12,6 @@ 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 <algorithm>
-
 #include "paddle/fluid/framework/infershape_utils.h"
 #include "paddle/fluid/framework/op_registry.h"
 #include "paddle/phi/infermeta/unary.h"
@@ -58,15 +56,56 @@ class DiagV2OpMaker : public framework::OpProtoAndCheckerMaker {
   }
 };
 
+class DiagV2GradOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+  void InferShape(framework::InferShapeContext *ctx) const override {
+    OP_INOUT_CHECK(ctx->HasInput("X"), "X", "X", "DiagV2Grad");
+    OP_INOUT_CHECK(ctx->HasOutput(framework::GradVarName("X")), "Output",
+                   framework::GradVarName("X"), "DiagV2Grad");
+
+    ctx->SetOutputDim(framework::GradVarName("X"), ctx->GetInputDim("X"));
+  }
+
+ protected:
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext &ctx) const override {
+    return framework::OpKernelType(OperatorWithKernel::IndicateVarDataType(
+                                       ctx, framework::GradVarName("Out")),
+                                   ctx.GetPlace());
+  }
+};
+
+template <typename T>
+class DiagV2GradOpMaker : public framework::SingleGradOpMaker<T> {
+ public:
+  using framework::SingleGradOpMaker<T>::SingleGradOpMaker;
+
+ protected:
+  void Apply(GradOpPtr<T> grad_op) const override {
+    grad_op->SetType("diag_v2_grad");
+    grad_op->SetInput("X", this->Input("X"));
+    grad_op->SetInput(framework::GradVarName("Out"), this->OutputGrad("Out"));
+    grad_op->SetOutput(framework::GradVarName("X"), this->InputGrad("X"));
+    grad_op->SetAttrMap(this->Attrs());
+  }
+};
+
+DECLARE_NO_NEED_BUFFER_VARS_INFERER(DiagGradV2NoNeedBufferVarsInferer, "X");
+
 }  // namespace operators
 }  // namespace paddle
 
 namespace ops = paddle::operators;
+
 DECLARE_INFER_SHAPE_FUNCTOR(diag_v2, DiagInferShapeFunctor,
                             PD_INFER_META(phi::DiagInferMeta));
 
-REGISTER_OPERATOR(
-    diag_v2, ops::DiagV2Op, ops::DiagV2OpMaker,
-    paddle::framework::EmptyGradOpMaker<paddle::framework::OpDesc>,
-    paddle::framework::EmptyGradOpMaker<paddle::imperative::OpBase>,
-    DiagInferShapeFunctor);
+REGISTER_OPERATOR(diag_v2, ops::DiagV2Op, ops::DiagV2OpMaker,
+                  ops::DiagV2GradOpMaker<paddle::framework::OpDesc>,
+                  ops::DiagV2GradOpMaker<paddle::imperative::OpBase>,
+                  DiagInferShapeFunctor);
+
+REGISTER_OPERATOR(diag_v2_grad, ops::DiagV2GradOp,
+                  ops::DiagGradV2NoNeedBufferVarsInferer);

paddle/phi/kernels/cpu/diag_grad_kernel.cc

+// 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/diag_grad_kernel.h"
+#include "paddle/phi/backends/cpu/cpu_context.h"
+#include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/kernels/funcs/diag_functor.h"
+#include "paddle/phi/kernels/funcs/math_function.h"
+
+namespace phi {
+
+template <typename T, typename Context>
+void DiagGradKernel(const Context& dev_ctx,
+                    const DenseTensor& x,
+                    const DenseTensor& out_grad,
+                    int offset,
+                    DenseTensor* x_grad) {
+  T* dx_data = dev_ctx.template Alloc<T>(x_grad);
+  const T* dout_data = out_grad.data<T>();
+  auto dx_dims = x_grad->dims();
+  auto dout_dims = out_grad.dims();
+
+  if (dx_dims.size() == 1) {
+    auto dx_length = dx_dims[0];
+    int dx_stride = phi::funcs::ComputeStride(0, dx_dims);
+
+    auto dout_stride_0 = phi::funcs::ComputeStride(0, dout_dims);
+    auto dout_stride_1 = phi::funcs::ComputeStride(1, dout_dims);
+    dout_data +=
+        (offset >= 0 ? offset * dout_stride_1 : -offset * dout_stride_0);
+
+    for (int i = 0; i < dx_length; i++) {
+      dx_data[i * dx_stride] = dout_data[i * (dout_stride_0 + dout_stride_1)];
+    }
+  } else {
+    phi::funcs::SetConstant<Context, T> set_padding_value;
+    set_padding_value(dev_ctx, x_grad, static_cast<T>(0));
+
+    int dx_stride_0 = phi::funcs::ComputeStride(0, dx_dims);
+    int dx_stride_1 = phi::funcs::ComputeStride(1, dx_dims);
+    auto dout_stride_0 = phi::funcs::ComputeStride(0, dout_dims);
+    dx_data += (offset >= 0 ? offset * dx_stride_1 : -offset * dx_stride_0);
+
+    auto dout_length = dout_dims[0];
+    for (int i = 0; i < dout_length; i++) {
+      dx_data[i * (dx_stride_0 + dx_stride_1)] = dout_data[i * dout_stride_0];
+    }
+  }
+}
+
+}  // namespace phi
+
+PD_REGISTER_KERNEL(diag_grad,
+                   CPU,
+                   ALL_LAYOUT,
+                   phi::DiagGradKernel,
+                   phi::dtype::float16,
+                   int,
+                   int64_t,
+                   float,
+                   double) {}

paddle/phi/kernels/cpu/diag_kernel.cc

@@ -62,5 +62,12 @@ void DiagKernel(const Context& dev_ctx,
 
 }  // namespace phi
 
-PD_REGISTER_KERNEL(
-    diag, CPU, ALL_LAYOUT, phi::DiagKernel, int, float, double, int64_t) {}
+PD_REGISTER_KERNEL(diag,
+                   CPU,
+                   ALL_LAYOUT,
+                   phi::DiagKernel,
+                   phi::dtype::float16,
+                   int,
+                   float,
+                   double,
+                   int64_t) {}

paddle/phi/kernels/diag_grad_kernel.h

+// 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.
+
+#pragma once
+
+#include "paddle/phi/core/dense_tensor.h"
+
+namespace phi {
+
+template <typename T, typename Context>
+void DiagGradKernel(const Context& dev_ctx,
+                    const DenseTensor& x,
+                    const DenseTensor& out_grad,
+                    int offset,
+                    DenseTensor* x_grad);
+
+}  // namespace phi

paddle/phi/kernels/gpu/diag_grad_kernel.cu

+// 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/diag_kernel.h"
+
+#include "paddle/phi/backends/gpu/gpu_context.h"
+#include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/kernels/funcs/diag_functor.h"
+#include "paddle/phi/kernels/funcs/math_function.h"
+
+namespace phi {
+
+// Extract the diagonal of a matrix 'dout' to a matrix 'dx'
+template <typename T>
+__global__ void ExtractDiagonalKernel(const T* dout,
+                                      T* dx,
+                                      std::ptrdiff_t start,
+                                      std::ptrdiff_t dx_length,
+                                      const std::ptrdiff_t sumStride,
+                                      const std::ptrdiff_t xStride) {
+  for (std::ptrdiff_t idx = blockIdx.x * blockDim.x + threadIdx.x;
+       idx < dx_length;
+       idx += gridDim.x * blockDim.x) {
+    const std::ptrdiff_t outOffset = start + sumStride * idx;
+    dx[xStride * idx] = dout[outOffset];
+  }
+}
+
+// Paste a vector 'dout' to the diagonal of a matrix 'dx'
+template <typename T>
+__global__ void PasteDiagonalKernel(const T* dout,
+                                    T* dx,
+                                    std::ptrdiff_t start,
+                                    std::ptrdiff_t size,
+                                    const std::ptrdiff_t sumStride,
+                                    const std::ptrdiff_t outStride) {
+  for (std::ptrdiff_t idx = blockIdx.x * blockDim.x + threadIdx.x; idx < size;
+       idx += gridDim.x * blockDim.x) {
+    std::ptrdiff_t xOffset = start + sumStride * idx;
+    dx[xOffset] = dout[outStride * idx];
+  }
+}
+
+template <typename T, typename Context>
+void DiagGradKernel(const Context& dev_ctx,
+                    const DenseTensor& x,
+                    const DenseTensor& out_grad,
+                    int offset,
+                    DenseTensor* x_grad) {
+  T* dx_data = dev_ctx.template Alloc<T>(x_grad);
+  auto* dout_data = out_grad.data<T>();
+  auto dx_dims = x_grad->dims();
+  auto dout_dims = out_grad.dims();
+
+  auto GetBlockGridSize = [&dev_ctx](int64_t size) {
+    const int64_t block_size =
+        std::min(size, static_cast<int64_t>(dev_ctx.GetMaxThreadsPerBlock()));
+    int64_t max_threads = dev_ctx.GetMaxPhysicalThreadCount();
+    const int64_t max_blocks =
+        std::max(((max_threads - 1) / block_size + 1), static_cast<int64_t>(1));
+    const int64_t grid_size =
+        std::min(max_blocks, (size + block_size - 1) / block_size);
+    return std::tuple<int64_t, int64_t>{block_size, grid_size};
+  };
+
+  if (dx_dims.size() == 1) {
+    auto dx_length = dx_dims[0];
+    auto size = (offset > 0) ? dx_length + offset : dx_length - offset;
+    int dx_stride = phi::funcs::ComputeStride(0, dx_dims);
+    if (size > 0) {
+      auto dout_stride_0 = phi::funcs::ComputeStride(0, dout_dims);
+      auto dout_stride_1 = phi::funcs::ComputeStride(1, dout_dims);
+      auto start =
+          (offset >= 0 ? offset * dout_stride_1 : -offset * dout_stride_0);
+
+      std::tuple<int64_t, int64_t> block_grid_size = GetBlockGridSize(size);
+      ExtractDiagonalKernel<T><<<std::get<1>(block_grid_size),
+                                 std::get<0>(block_grid_size),
+                                 0,
+                                 dev_ctx.stream()>>>(
+          dout_data,
+          dx_data,
+          start,
+          dx_length,
+          dout_stride_0 + dout_stride_1,
+          dx_stride);
+    }
+  } else {
+    phi::funcs::SetConstant<Context, T> set_padding_value;
+    set_padding_value(dev_ctx, x_grad, static_cast<T>(0));
+
+    int dx_stride_0 = phi::funcs::ComputeStride(0, dx_dims);
+    int dx_stride_1 = phi::funcs::ComputeStride(1, dx_dims);
+    int64_t size;
+    if (offset > 0) {
+      size = std::min(dx_dims[0], dx_dims[1] - offset);
+    } else {
+      size = std::min(dx_dims[0] + offset, dx_dims[1]);
+    }
+
+    if (size > 0) {
+      auto start = (offset >= 0 ? offset * dx_stride_1 : -offset * dx_stride_0);
+      auto dout_stride_0 = phi::funcs::ComputeStride(0, dout_dims);
+      std::tuple<int64_t, int64_t> block_grid_size = GetBlockGridSize(size);
+      PasteDiagonalKernel<T><<<std::get<1>(block_grid_size),
+                               std::get<0>(block_grid_size),
+                               0,
+                               dev_ctx.stream()>>>(dout_data,
+                                                   dx_data,
+                                                   start,
+                                                   size,
+                                                   dx_stride_0 + dx_stride_1,
+                                                   dout_stride_0);
+    }
+  }
+}
+
+}  // namespace phi
+
+PD_REGISTER_KERNEL(diag_grad,
+                   GPU,
+                   ALL_LAYOUT,
+                   phi::DiagGradKernel,
+                   phi::dtype::float16,
+                   int,
+                   int64_t,
+                   float,
+                   double) {}

paddle/phi/kernels/gpu/diag_kernel.cu

@@ -130,5 +130,12 @@ void DiagKernel(const Context& dev_ctx,
 
 }  // namespace phi
 
-PD_REGISTER_KERNEL(
-    diag, GPU, ALL_LAYOUT, phi::DiagKernel, int, int64_t, float, double) {}
+PD_REGISTER_KERNEL(diag,
+                   GPU,
+                   ALL_LAYOUT,
+                   phi::DiagKernel,
+                   phi::dtype::float16,
+                   int,
+                   int64_t,
+                   float,
+                   double) {}

paddle/phi/ops/compat/diag_sig.cc

@@ -20,8 +20,15 @@ KernelSignature DiagOpArgumentMapping(const ArgumentMappingContext& ctx) {
   return KernelSignature("diag", {"X"}, {"offset", "padding_value"}, {"Out"});
 }
 
+KernelSignature DiagGradOpArgumentMapping(const ArgumentMappingContext& ctx) {
+  return KernelSignature(
+      "diag_grad", {"X", GradVarName("Out")}, {"offset"}, {GradVarName("X")});
+}
+
 }  // namespace phi
 
 PD_REGISTER_BASE_KERNEL_NAME(diag_v2, diag);
+PD_REGISTER_BASE_KERNEL_NAME(diag_v2_grad, diag_grad);
 
 PD_REGISTER_ARG_MAPPING_FN(diag_v2, phi::DiagOpArgumentMapping);
+PD_REGISTER_ARG_MAPPING_FN(diag_v2_grad, phi::DiagGradOpArgumentMapping);

python/paddle/fluid/tests/unittests/test_diag_v2.py

@@ -44,6 +44,10 @@ class TestDiagV2Op(OpTest):
         paddle.enable_static()
         self.check_output(check_eager=True)
 
+    def test_check_grad(self):
+        paddle.enable_static()
+        self.check_grad(['X'], 'Out', check_eager=True)
+
     def init_config(self):
         pass
 
@@ -62,14 +66,14 @@ class TestDiagV2OpCase2(TestDiagV2Op):
 
 class TestDiagV2OpCase3(TestDiagV2Op):
     def init_config(self):
-        self.x = np.random.randint(-10, 10, size=(10, 10))
+        self.x = np.random.randint(-10, 10, size=(10, 10)).astype("float64")
         self.out = np.diag(self.x, self.offset)
 
 
 class TestDiagV2OpCase4(TestDiagV2Op):
     def init_config(self):
         self.x = np.random.rand(100)
-        self.padding_value = 8
+        self.padding_value = 2
         n = self.x.size
         self.out = self.padding_value * np.ones((n, n)) + np.diag(
             self.x, self.offset) - np.diag(self.padding_value * np.ones(n))