Benchmarks for dynamic autodiff.

This patch is from Clement Courbet. courbet@google.com

Change-Id: I886390663644733bfa5b7b52b0c883079e793726

internal/ceres/autodiff_benchmarks/autodiff_benchmarks.cc

@@ -30,6 +30,7 @@
 
 #include <memory>
 #include <random>
+#include <utility>
 
 #include "benchmark/benchmark.h"
 #include "ceres/autodiff_benchmarks/brdf_cost_function.h"
@@ -41,30 +42,32 @@
 #include "ceres/ceres.h"
 
 namespace ceres {
-namespace internal {
 
-// If we want to use functors with both operator() and an Evaluate() method
-// with AutoDiff then this wrapper class here has to be used. Autodiff doesn't
-// support functors that have an Evaluate() function.
-//
-// CostFunctionToFunctor hides the Evaluate() function, because it doesn't
-// derive from CostFunction. Autodiff sees it as a simple functor and will use
-// the operator() as expected.
-template <typename CostFunction>
-struct CostFunctionToFunctor {
-    template <typename... _Args>
-    explicit CostFunctionToFunctor(_Args&&... __args)
-        : cost_function(std::forward<_Args>(__args)...) {}
-
-    template <typename... _Args>
-    inline bool operator()(_Args&&... __args) const {
-        return cost_function(std::forward<_Args>(__args)...);
-    }
-
-    CostFunction cost_function;
-};
+enum Dynamic { kNotDynamic, kDynamic };
+
+// Transforms a static functor into a dynamic one.
+template <typename CostFunctionType, int kNumParameterBlocks>
+class ToDynamic {
+ public:
+  template <typename... _Args>
+  explicit ToDynamic(_Args&&... __args)
+      : cost_function_(std::forward<_Args>(__args)...) {}
+
+  template <typename T>
+  bool operator()(const T* const* parameters, T* residuals) const {
+    return Apply(parameters, residuals,
+                 std::make_index_sequence<kNumParameterBlocks>());
+  }
+
+ private:
+  template <typename T, size_t... Indices>
+  bool Apply(const T* const* parameters, T* residuals,
+             std::index_sequence<Indices...>) const {
+    return cost_function_(parameters[Indices]..., residuals);
+  }
 
-}  // namespace internal
+  CostFunctionType cost_function_;
+};
 
 template <int kParameterBlockSize>
 static void BM_ConstantAnalytic(benchmark::State& state) {
@@ -79,14 +82,59 @@ static void BM_ConstantAnalytic(benchmark::State& state) {
   double* jacobians[] = {jacobian_values.data()};
 
   std::unique_ptr<ceres::CostFunction> cost_function(
-      new ConstantCostFunction<kParameterBlockSize>());
+      new AnalyticConstantCostFunction<kParameterBlockSize>());
 
   for (auto _ : state) {
     cost_function->Evaluate(parameters, residuals.data(), jacobians);
   }
 }
 
-template <int kParameterBlockSize>
+// Helpers for CostFunctionFactory.
+template <typename DynamicCostFunctionType>
+void AddParameterBlocks(DynamicCostFunctionType*) {}
+
+template <int HeadN, int... TailNs, typename DynamicCostFunctionType>
+void AddParameterBlocks(DynamicCostFunctionType* dynamic_function) {
+  dynamic_function->AddParameterBlock(HeadN);
+  AddParameterBlocks<TailNs...>(dynamic_function);
+}
+
+// Creates an autodiff cost function wrapping `CostFunctor`, with
+// `kNumResiduals` residuals and parameter blocks with sized `Ns..`.
+// Depending on `kIsDynamic`, either a static or dynamic cost function is
+// created.
+// `args` are forwarded to the `CostFunctor` constructor.
+template <Dynamic kIsDynamic>
+struct CostFunctionFactory {};
+
+template <>
+struct CostFunctionFactory<kNotDynamic> {
+  template <typename CostFunctor, int kNumResiduals, int... Ns,
+            typename... Args>
+  static std::unique_ptr<ceres::CostFunction> Create(Args&&... args) {
+    return std::make_unique<
+        ceres::AutoDiffCostFunction<CostFunctor, kNumResiduals, Ns...>>(
+        new CostFunctor(std::forward<Args>(args)...));
+  }
+};
+
+template <>
+struct CostFunctionFactory<kDynamic> {
+  template <typename CostFunctor, int kNumResiduals, int... Ns,
+            typename... Args>
+  static std::unique_ptr<ceres::CostFunction> Create(Args&&... args) {
+    constexpr const int kNumParameterBlocks = sizeof...(Ns);
+    auto dynamic_function = std::make_unique<ceres::DynamicAutoDiffCostFunction<
+        ToDynamic<CostFunctor, kNumParameterBlocks>>>(
+        new ToDynamic<CostFunctor, kNumParameterBlocks>(
+            std::forward<Args>(args)...));
+    dynamic_function->SetNumResiduals(kNumResiduals);
+    AddParameterBlocks<Ns...>(dynamic_function.get());
+    return dynamic_function;
+  }
+};
+
+template <int kParameterBlockSize, Dynamic kIsDynamic>
 static void BM_ConstantAutodiff(benchmark::State& state) {
   constexpr int num_residuals = 1;
   std::array<double, kParameterBlockSize> parameters_values;
@@ -98,11 +146,9 @@ static void BM_ConstantAutodiff(benchmark::State& state) {
   std::array<double, num_residuals * kParameterBlockSize> jacobian_values;
   double* jacobians[] = {jacobian_values.data()};
 
-  using AutoDiffFunctor = ceres::internal::CostFunctionToFunctor<
-      ConstantCostFunction<kParameterBlockSize>>;
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<AutoDiffFunctor, 1, kParameterBlockSize>(
-          new AutoDiffFunctor()));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<
+          ConstantCostFunction<kParameterBlockSize>, 1, 1>();
 
   for (auto _ : state) {
     cost_function->Evaluate(parameters, residuals.data(), jacobians);
@@ -110,24 +156,29 @@ static void BM_ConstantAutodiff(benchmark::State& state) {
 }
 
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 1);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 1);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 1, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 1, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 10);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 10);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 10, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 10, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 20);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 20);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 20, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 20, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 30);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 30);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 30, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 30, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 40);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 40);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 40, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 40, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 50);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 50);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 50, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 50, kDynamic);
 BENCHMARK_TEMPLATE(BM_ConstantAnalytic, 60);
-BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 60);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 60, kNotDynamic);
+BENCHMARK_TEMPLATE(BM_ConstantAutodiff, 60, kDynamic);
 
+template <Dynamic kIsDynamic>
 static void BM_Linear1AutoDiff(benchmark::State& state) {
-  using FunctorType =
-      ceres::internal::CostFunctionToFunctor<Linear1CostFunction>;
-
   double parameter_block1[] = {1.};
   double* parameters[] = {parameter_block1};
 
@@ -135,20 +186,20 @@ static void BM_Linear1AutoDiff(benchmark::State& state) {
   double residuals[1];
   double* jacobians[] = {jacobian1};
 
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType, 1, 1>(new FunctorType()));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<Linear1CostFunction, 1,
+                                                       1>();
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
-BENCHMARK(BM_Linear1AutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Linear1AutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Linear1AutoDiff, kDynamic)->Arg(0)->Arg(1);
 
+template <Dynamic kIsDynamic>
 static void BM_Linear10AutoDiff(benchmark::State& state) {
-  using FunctorType =
-      ceres::internal::CostFunctionToFunctor<Linear10CostFunction>;
-
   double parameter_block1[] = {1., 2., 3., 4., 5., 6., 7., 8., 9., 10.};
   double* parameters[] = {parameter_block1};
 
@@ -156,15 +207,17 @@ static void BM_Linear10AutoDiff(benchmark::State& state) {
   double residuals[10];
   double* jacobians[] = {jacobian1};
 
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType, 10, 10>(new FunctorType()));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<Linear10CostFunction, 10,
+                                                       10>();
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
-BENCHMARK(BM_Linear10AutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Linear10AutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Linear10AutoDiff, kDynamic)->Arg(0)->Arg(1);
 
 // From the NIST problem collection.
 struct Rat43CostFunctor {
@@ -180,11 +233,14 @@ struct Rat43CostFunctor {
     return true;
   }
 
+  static constexpr int kNumParameterBlocks = 1;
+
  private:
   const double x_;
   const double y_;
 };
 
+template <Dynamic kIsDynamic>
 static void BM_Rat43AutoDiff(benchmark::State& state) {
   double parameter_block1[] = {1., 2., 3., 4.};
   double* parameters[] = {parameter_block1};
@@ -194,21 +250,20 @@ static void BM_Rat43AutoDiff(benchmark::State& state) {
   double* jacobians[] = {jacobian1};
   const double x = 0.2;
   const double y = 0.3;
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<Rat43CostFunctor, 1, 4>(
-          new Rat43CostFunctor(x, y)));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<Rat43CostFunctor, 1, 4>(
+          x, y);
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, &residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, &residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
-BENCHMARK(BM_Rat43AutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Rat43AutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_Rat43AutoDiff, kDynamic)->Arg(0)->Arg(1);
 
+template <Dynamic kIsDynamic>
 static void BM_SnavelyReprojectionAutoDiff(benchmark::State& state) {
-  using FunctorType =
-      ceres::internal::CostFunctionToFunctor<SnavelyReprojectionError>;
-
   double parameter_block1[] = {1., 2., 3., 4., 5., 6., 7., 8., 9.};
   double parameter_block2[] = {1., 2., 3.};
   double* parameters[] = {parameter_block1, parameter_block2};
@@ -220,18 +275,20 @@ static void BM_SnavelyReprojectionAutoDiff(benchmark::State& state) {
 
   const double x = 0.2;
   const double y = 0.3;
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType, 2, 9, 3>(
-          new FunctorType(x, y)));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<SnavelyReprojectionError,
+                                                       2, 9, 3>(x, y);
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
 
-BENCHMARK(BM_SnavelyReprojectionAutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_SnavelyReprojectionAutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_SnavelyReprojectionAutoDiff, kDynamic)->Arg(0)->Arg(1);
 
+template <Dynamic kIsDynamic>
 static void BM_PhotometricAutoDiff(benchmark::State& state) {
   constexpr int PATCH_SIZE = 8;
 
@@ -280,22 +337,22 @@ static void BM_PhotometricAutoDiff(benchmark::State& state) {
   FunctorType::Intrinsics intrinsics;
   intrinsics << 128, 128, 1, -1, 0.5, 0.5;
 
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType,
-                                      FunctorType::PATCH_SIZE,
-                                      FunctorType::POSE_SIZE,
-                                      FunctorType::POSE_SIZE,
-                                      FunctorType::POINT_SIZE>(new FunctorType(
-          intensities_host, bearings_host, image_target, intrinsics)));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<
+          FunctorType, FunctorType::PATCH_SIZE, FunctorType::POSE_SIZE,
+          FunctorType::POSE_SIZE, FunctorType::POINT_SIZE>(
+          intensities_host, bearings_host, image_target, intrinsics);
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
 
-BENCHMARK(BM_PhotometricAutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_PhotometricAutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_PhotometricAutoDiff, kDynamic)->Arg(0)->Arg(1);
 
+template <Dynamic kIsDynamic>
 static void BM_RelativePoseAutoDiff(benchmark::State& state) {
   using FunctorType = RelativePoseError;
 
@@ -314,20 +371,22 @@ static void BM_RelativePoseAutoDiff(benchmark::State& state) {
   Eigen::Quaterniond q_i_j = Eigen::Quaterniond(1, 2, 3, 4).normalized();
   Eigen::Vector3d t_i_j(1, 2, 3);
 
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType, 6, 7, 7>(
-          new FunctorType(q_i_j, t_i_j)));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<FunctorType, 6, 7, 7>(
+          q_i_j, t_i_j);
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
 
-BENCHMARK(BM_RelativePoseAutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_RelativePoseAutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_RelativePoseAutoDiff, kDynamic)->Arg(0)->Arg(1);
 
+template <Dynamic kIsDynamic>
 static void BM_BrdfAutoDiff(benchmark::State& state) {
-  using FunctorType = ceres::internal::CostFunctionToFunctor<Brdf>;
+  using FunctorType = Brdf;
 
   double material[] = {1., 2., 3., 4., 5., 6., 7., 8., 9., 10.};
   auto c = Eigen::Vector3d(0.1, 0.2, 0.3);
@@ -337,32 +396,29 @@ static void BM_BrdfAutoDiff(benchmark::State& state) {
   auto x = Eigen::Vector3d(0.5, 0.7, -0.1).normalized();
   auto y = Eigen::Vector3d(0.2, -0.2, -0.2).normalized();
 
-  double* parameters[7] = {
-      material, c.data(), n.data(), v.data(), l.data(), x.data(), y.data()};
+  double* parameters[7] = {material, c.data(), n.data(), v.data(),
+                           l.data(), x.data(), y.data()};
 
   double jacobian[(10 + 6 * 3) * 3];
   double residuals[3];
   double* jacobians[7] = {
-      jacobian + 0,
-      jacobian + 10 * 3,
-      jacobian + 13 * 3,
-      jacobian + 16 * 3,
-      jacobian + 19 * 3,
-      jacobian + 22 * 3,
+      jacobian + 0,      jacobian + 10 * 3, jacobian + 13 * 3,
+      jacobian + 16 * 3, jacobian + 19 * 3, jacobian + 22 * 3,
       jacobian + 25 * 3,
   };
 
-  std::unique_ptr<ceres::CostFunction> cost_function(
-      new ceres::AutoDiffCostFunction<FunctorType, 3, 10, 3, 3, 3, 3, 3, 3>(
-          new FunctorType));
+  std::unique_ptr<ceres::CostFunction> cost_function =
+      CostFunctionFactory<kIsDynamic>::template Create<FunctorType, 3, 10, 3, 3,
+                                                       3, 3, 3, 3>();
 
   for (auto _ : state) {
-    cost_function->Evaluate(
-        parameters, residuals, state.range(0) ? jacobians : nullptr);
+    cost_function->Evaluate(parameters, residuals,
+                            state.range(0) ? jacobians : nullptr);
   }
 }
 
-BENCHMARK(BM_BrdfAutoDiff)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_BrdfAutoDiff, kNotDynamic)->Arg(0)->Arg(1);
+BENCHMARK_TEMPLATE(BM_BrdfAutoDiff, kDynamic)->Arg(0)->Arg(1);
 
 }  // namespace ceres
 

internal/ceres/autodiff_benchmarks/constant_cost_function.h

@@ -37,14 +37,17 @@
 namespace ceres {
 
 template <int kParameterBlockSize>
-struct ConstantCostFunction
-    : public ceres::SizedCostFunction<1, kParameterBlockSize> {
+struct ConstantCostFunction {
   template <typename T>
   inline bool operator()(const T* const x, T* residuals) const {
     residuals[0] = T(5);
     return true;
   }
+};
 
+template <int kParameterBlockSize>
+struct AnalyticConstantCostFunction
+    : public ceres::SizedCostFunction<1, kParameterBlockSize> {
   virtual bool Evaluate(double const* const* parameters,
                         double* residuals,
                         double** jacobians) const {