sorted_k_distances.h File Reference

#include <Eigen/Core>
#include <algorithm>
#include <array>
#include <cstring>
#include <limits>

Include dependency graph for sorted_k_distances.h:

This graph shows which files directly or indirectly include this file:

Classes
class	occ::descriptors::SortedKDistances< K >

Namespaces
namespace	occ
namespace	occ::descriptors

Macros
#define	HAS_NEON   0
	High-performance helper class to maintain k smallest distances in sorted order.
#define	HAS_AVX2   0

Typedefs
using	occ::descriptors::SortedK1Distance = SortedKDistances< 1 >
using	occ::descriptors::SortedK5Distances = SortedKDistances< 5 >
using	occ::descriptors::SortedK10Distances = SortedKDistances< 10 >
using	occ::descriptors::SortedK20Distances = SortedKDistances< 20 >
using	occ::descriptors::SortedK50Distances = SortedKDistances< 50 >
using	occ::descriptors::SortedK100Distances = SortedKDistances< 100 >

Macro Definition Documentation

HAS_AVX2

#define HAS_AVX2   0

HAS_NEON

#define HAS_NEON   0

High-performance helper class to maintain k smallest distances in sorted order.

Template parameter K allows compile-time optimizations and stack allocation Automatically uses NEON on ARM or AVX2 on x86

Optimizations:

• No dynamic allocation - everything on stack
• SIMD operations for comparisons and shifts (NEON or AVX2)
• Batch insertion support for processing multiple values at once
• Binary search for larger k values
• Compile-time loop unrolling
• Branchless operations where possible