Files
ceres-solver/internal/ceres/program.h
T
Dmitriy Korchemkin b158515089 Parallel operations on vectors
Main focus of this change is to parallelize remaining operations (most of them
are operations on vectors) in code-path utilized with iterative Schur
complement.

Parallelization is handled using lazy evaluation of Eigen expressions.

On linux pc with intel 8176 processor parallelization of vector operations has
the following effect:

Running ./bin/parallel_vector_operations_benchmark
Run on (112 X 3200.32 MHz CPU s)
CPU Caches:
  L1 Data 32 KiB (x56)
  L1 Instruction 32 KiB (x56)
  L2 Unified 1024 KiB (x56)
  L3 Unified 39424 KiB (x2)
Load Average: 3.30, 8.41, 11.82
-----------------------------------
Benchmark                      Time
-----------------------------------
SetZero                 10009532 ns
SetZeroParallel/1       10024139 ns
...
SetZeroParallel/16        877606 ns

Negate                   4978856 ns
NegateParallel/1         5145413 ns
...
NegateParallel/16         721823 ns

Assign                  10731408 ns
AssignParallel/1        10749944 ns
...
AssignParallel/16        1829381 ns

D2X                     15214399 ns
D2XParallel/1           15623245 ns
...
D2XParallel/16           2687060 ns

DivideSqrt               8220050 ns
DivideSqrtParallel/1     9088467 ns
...
DivideSqrtParallel/16     905569 ns

Clamp                    3502010 ns
ClampParallel/1          4507897 ns
...
ClampParallel/16          759576 ns

Norm                     4426782 ns
NormParallel/1           4442805 ns
...
NormParallel/16           430290 ns

Dot                      9023276 ns
DotParallel/1            9031304 ns
...
DotParallel/16           1157267 ns

Axpby                   14608289 ns
AxpbyParallel/1         14570825 ns
...
AxpbyParallel/16         2672220 ns
-----------------------------------

Multi-threading of vector operations in ISC and program evaluation results into
the following improvement:

Running ./bin/evaluation_benchmark
--------------------------------------------------------------------------------------
Benchmark                                                               this   2fd81de
--------------------------------------------------------------------------------------
Residuals<problem-13682-4456117-pre.txt>/1                           4136 ms   4292 ms
Residuals<problem-13682-4456117-pre.txt>/2                           2919 ms   2670 ms
Residuals<problem-13682-4456117-pre.txt>/4                           2065 ms   2198 ms
Residuals<problem-13682-4456117-pre.txt>/8                           1458 ms   1609 ms
Residuals<problem-13682-4456117-pre.txt>/16                          1152 ms   1227 ms

ResidualsAndJacobian<problem-13682-4456117-pre.txt>/1               19759 ms  20084 ms
ResidualsAndJacobian<problem-13682-4456117-pre.txt>/2               10921 ms  10977 ms
ResidualsAndJacobian<problem-13682-4456117-pre.txt>/4                6220 ms   6941 ms
ResidualsAndJacobian<problem-13682-4456117-pre.txt>/8                3490 ms   4398 ms
ResidualsAndJacobian<problem-13682-4456117-pre.txt>/16               2277 ms   3172 ms

Plus<problem-13682-4456117-pre.txt>/1                                 339 ms    322 ms
Plus<problem-13682-4456117-pre.txt>/2                                 220 ms
Plus<problem-13682-4456117-pre.txt>/4                                 128 ms
Plus<problem-13682-4456117-pre.txt>/8                                78.0 ms
Plus<problem-13682-4456117-pre.txt>/16                               49.8 ms

ISCRightMultiplyAndAccumulate<problem-13682-4456117-pre.txt>/1       2434 ms   2478 ms
ISCRightMultiplyAndAccumulate<problem-13682-4456117-pre.txt>/2       2706 ms   2688 ms
ISCRightMultiplyAndAccumulate<problem-13682-4456117-pre.txt>/4       1430 ms   1548 ms
ISCRightMultiplyAndAccumulate<problem-13682-4456117-pre.txt>/8        742 ms    883 ms
ISCRightMultiplyAndAccumulate<problem-13682-4456117-pre.txt>/16       438 ms    555 ms

ISCRightMultiplyAndAccumulateDiag<problem-13682-4456117-pre.txt>/1   2438 ms   2481 ms
ISCRightMultiplyAndAccumulateDiag<problem-13682-4456117-pre.txt>/2   2565 ms   2790 ms
ISCRightMultiplyAndAccumulateDiag<problem-13682-4456117-pre.txt>/4   1434 ms   1551 ms
ISCRightMultiplyAndAccumulateDiag<problem-13682-4456117-pre.txt>/8    765 ms    892 ms
ISCRightMultiplyAndAccumulateDiag<problem-13682-4456117-pre.txt>/16   435 ms    559 ms

JacobianSquaredColumnNorm<problem-13682-4456117-pre.txt>/1           1278 ms
JacobianSquaredColumnNorm<problem-13682-4456117-pre.txt>/2           1555 ms
JacobianSquaredColumnNorm<problem-13682-4456117-pre.txt>/4            833 ms
JacobianSquaredColumnNorm<problem-13682-4456117-pre.txt>/8            459 ms
JacobianSquaredColumnNorm<problem-13682-4456117-pre.txt>/16           250 ms

JacobianScaleColumns<problem-13682-4456117-pre.txt>/1                1468 ms
JacobianScaleColumns<problem-13682-4456117-pre.txt>/2                1871 ms
JacobianScaleColumns<problem-13682-4456117-pre.txt>/4                 957 ms
JacobianScaleColumns<problem-13682-4456117-pre.txt>/8                 528 ms
JacobianScaleColumns<problem-13682-4456117-pre.txt>/16                294 ms

End-to-end improvements with bundle_adjuster invoked with
./bin/bundle_adjuster --num_threads 28 --num_iterations 40 \
                      --linear_solver iterative_schur \
                      --preconditioner jacobi --input
---------------------------------------------
Problem                         this  2fd81de
---------------------------------------------
problem-13682-4456117-pre.txt  508.6    892.7
problem-1778-993923-pre.txt    763.8   1129.9
problem-1723-156502-pre.txt      6.3     14.4
problem-356-226730-pre.txt      76.3    116.2
problem-257-65132-pre.txt       38.6     52.0

Change-Id: Ie31cc5015f13fa479c16ffb5ce48c9b880990d49
2022-12-17 02:52:27 +03:00

202 lines
8.2 KiB
C++

// Ceres Solver - A fast non-linear least squares minimizer
// Copyright 2015 Google Inc. All rights reserved.
// http://ceres-solver.org/
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
//
// * Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
// * Neither the name of Google Inc. nor the names of its contributors may be
// used to endorse or promote products derived from this software without
// specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
// POSSIBILITY OF SUCH DAMAGE.
//
// Author: keir@google.com (Keir Mierle)
#ifndef CERES_INTERNAL_PROGRAM_H_
#define CERES_INTERNAL_PROGRAM_H_
#include <memory>
#include <set>
#include <string>
#include <vector>
#include "ceres/evaluation_callback.h"
#include "ceres/internal/disable_warnings.h"
#include "ceres/internal/export.h"
namespace ceres::internal {
class ParameterBlock;
class ProblemImpl;
class ResidualBlock;
class TripletSparseMatrix;
class ContextImpl;
// A nonlinear least squares optimization problem. This is different from the
// similarly-named "Problem" object, which offers a mutation interface for
// adding and modifying parameters and residuals. The Program contains the core
// part of the Problem, which is the parameters and the residuals, stored in a
// particular ordering. The ordering is critical, since it defines the mapping
// between (residual, parameter) pairs and a position in the jacobian of the
// objective function. Various parts of Ceres transform one Program into
// another; for example, the first stage of solving involves stripping all
// constant parameters and residuals. This is in contrast with Problem, which is
// not built for transformation.
class CERES_NO_EXPORT Program {
public:
// The ordered parameter and residual blocks for the program.
const std::vector<ParameterBlock*>& parameter_blocks() const;
const std::vector<ResidualBlock*>& residual_blocks() const;
std::vector<ParameterBlock*>* mutable_parameter_blocks();
std::vector<ResidualBlock*>* mutable_residual_blocks();
EvaluationCallback* mutable_evaluation_callback();
// Serialize to/from the program and update states.
//
// NOTE: Setting the state of a parameter block can trigger the
// computation of the Jacobian of its manifold. If this computation fails for
// some reason, then this method returns false and the state of the parameter
// blocks cannot be trusted.
bool StateVectorToParameterBlocks(const double* state);
void ParameterBlocksToStateVector(double* state) const;
// Copy internal state to the user's parameters.
void CopyParameterBlockStateToUserState();
// Set the parameter block pointers to the user pointers. Since this
// runs parameter block set state internally, which may call manifold, this
// can fail. False is returned on failure.
bool SetParameterBlockStatePtrsToUserStatePtrs();
// Update a state vector for the program given a delta.
bool Plus(const double* state,
const double* delta,
double* state_plus_delta,
ContextImpl* context,
int num_threads) const;
// Set the parameter indices and offsets. This permits mapping backward
// from a ParameterBlock* to an index in the parameter_blocks() vector. For
// any parameter block p, after calling SetParameterOffsetsAndIndex(), it
// is true that
//
// parameter_blocks()[p->index()] == p
//
// If a parameter appears in a residual but not in the parameter block, then
// it will have an index of -1.
//
// This also updates p->state_offset() and p->delta_offset(), which are the
// position of the parameter in the state and delta vector respectively.
void SetParameterOffsetsAndIndex();
// Check if the internal state of the program (the indexing and the
// offsets) are correct.
bool IsValid() const;
bool ParameterBlocksAreFinite(std::string* message) const;
// Returns true if the program has any non-constant parameter blocks
// which have non-trivial bounds constraints.
bool IsBoundsConstrained() const;
// Returns false, if the program has any constant parameter blocks
// which are not feasible, or any variable parameter blocks which
// have a lower bound greater than or equal to the upper bound.
bool IsFeasible(std::string* message) const;
// Loop over each residual block and ensure that no two parameter
// blocks in the same residual block are part of
// parameter_blocks as that would violate the assumption that it
// is an independent set in the Hessian matrix.
bool IsParameterBlockSetIndependent(
const std::set<double*>& independent_set) const;
// Create a TripletSparseMatrix which contains the zero-one
// structure corresponding to the block sparsity of the transpose of
// the Jacobian matrix.
//
// start_residual_block which allows the user to ignore the first
// start_residual_block residuals.
std::unique_ptr<TripletSparseMatrix> CreateJacobianBlockSparsityTranspose(
int start_residual_block = 0) const;
// Create a copy of this program and removes constant parameter
// blocks and residual blocks with no varying parameter blocks while
// preserving their relative order.
//
// removed_parameter_blocks on exit will contain the list of
// parameter blocks that were removed.
//
// fixed_cost will be equal to the sum of the costs of the residual
// blocks that were removed.
//
// If there was a problem, then the function will return a nullptr
// pointer and error will contain a human readable description of
// the problem.
std::unique_ptr<Program> CreateReducedProgram(
std::vector<double*>* removed_parameter_blocks,
double* fixed_cost,
std::string* error) const;
// See problem.h for what these do.
int NumParameterBlocks() const;
int NumParameters() const;
int NumEffectiveParameters() const;
int NumResidualBlocks() const;
int NumResiduals() const;
int MaxScratchDoublesNeededForEvaluate() const;
int MaxDerivativesPerResidualBlock() const;
int MaxParametersPerResidualBlock() const;
int MaxResidualsPerResidualBlock() const;
// A human-readable dump of the parameter blocks for debugging.
// TODO(keir): If necessary, also dump the residual blocks.
std::string ToString() const;
private:
// Remove constant parameter blocks and residual blocks with no
// varying parameter blocks while preserving their relative order.
//
// removed_parameter_blocks on exit will contain the list of
// parameter blocks that were removed.
//
// fixed_cost will be equal to the sum of the costs of the residual
// blocks that were removed.
//
// If there was a problem, then the function will return false and
// error will contain a human readable description of the problem.
bool RemoveFixedBlocks(std::vector<double*>* removed_parameter_blocks,
double* fixed_cost,
std::string* message);
// The Program does not own the ParameterBlock or ResidualBlock objects.
std::vector<ParameterBlock*> parameter_blocks_;
std::vector<ResidualBlock*> residual_blocks_;
EvaluationCallback* evaluation_callback_ = nullptr;
friend class ProblemImpl;
};
} // namespace ceres::internal
#include "ceres/internal/reenable_warnings.h"
#endif // CERES_INTERNAL_PROGRAM_H_