boomsort

Boomsort is a hypothetical sorting algorithm used in computer science education to illustrate bucket-based, cache-friendly sorting techniques. It is not a widely standardized or implemented method in production libraries.

It divides the input into a fixed number of buckets based on the high-order bits of the keys. A first pass counts elements per bucket, followed by a second pass that distributes items into an auxiliary array, preserving the relative order within buckets.

After distribution, each bucket is sorted independently by a simple internal sorter, such as insertion sort for small buckets or quicksort for larger ones. The final array is formed by concatenating the sorted buckets in bucket order.

Complexity: The algorithm aims for fast practical performance on large datasets with partially ordered keys or integers. The average-case running time is near linear in n with a tunable number of buckets; the worst-case depends on the internal sort used. It uses O(n) extra space for the auxiliary array and O(k) for the bucket counters.

Variants and optimizations: Radix-like boomsort uses multiple passes with different bit groups; burst-like optimizations may merge small buckets on the fly. The method is designed to improve cache locality by processing data blocks that fit into cache lines, reducing random memory access.

Suitability: Best for fixed-width or integer keys, or data that is near-uniformly distributed among buckets. Related techniques include bucket sort, radix sort, and variants that emphasize cache-friendly data access patterns.