Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve Runtime by reshaping the DataAssociation #411

Open
wants to merge 1 commit into
base: tiziano/reduce_allocations
Choose a base branch
from
Open

Improve Runtime by reshaping the DataAssociation #411

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 6 additions & 17 deletions cpp/kiss_icp/core/Registration.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,8 +23,10 @@
#include "Registration.hpp"

#include <tbb/blocked_range.h>
#include <tbb/concurrent_vector.h>
#include <tbb/global_control.h>
#include <tbb/info.h>
#include <tbb/parallel_for.h>
#include <tbb/parallel_reduce.h>
#include <tbb/task_arena.h>

Expand All @@ -44,7 +46,7 @@ using Matrix3_6d = Eigen::Matrix<double, 3, 6>;
using Vector6d = Eigen::Matrix<double, 6, 1>;
} // namespace Eigen

using Correspondences = std::vector<std::pair<Eigen::Vector3d, Eigen::Vector3d>>;
using Correspondences = tbb::concurrent_vector<std::pair<Eigen::Vector3d, Eigen::Vector3d>>;
using LinearSystem = std::pair<Eigen::Matrix6d, Eigen::Vector6d>;

namespace {
Expand All @@ -61,30 +63,17 @@ Correspondences DataAssociation(const std::vector<Eigen::Vector3d> &points,
using points_iterator = std::vector<Eigen::Vector3d>::const_iterator;
Correspondences correspondences;
correspondences.reserve(points.size());
correspondences = tbb::parallel_reduce(
tbb::parallel_for(
// Range
tbb::blocked_range<points_iterator>{points.cbegin(), points.cend()},
// Identity
correspondences,
// 1st lambda: Parallel computation
[&](const tbb::blocked_range<points_iterator> &r, Correspondences res) -> Correspondences {
res.reserve(r.size());
[&](const tbb::blocked_range<points_iterator> &r) {
std::for_each(r.begin(), r.end(), [&](const auto &point) {
const auto &[closest_neighbor, distance] = voxel_map.GetClosestNeighbor(point);
if (distance < max_correspondance_distance) {
res.emplace_back(point, closest_neighbor);
correspondences.emplace_back(point, closest_neighbor);
}
});
return res;
},
// 2nd lambda: Parallel reduction
[](Correspondences a, const Correspondences &b) -> Correspondences {
a.insert(a.end(), //
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems to be the problematic operation. It might require more investigation, though. Anyway, I feel that this reduction was inappropriate for the task here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What do you mean by problematic? Regarding the runtime?

std::make_move_iterator(b.cbegin()), //
std::make_move_iterator(b.cend()));
return a;
});

return correspondences;
}

Expand Down
Loading