This article introduces guided K-merge, a novel approach for merging multiple sorted sequences without traditional helper data structures like priority queues. Instead, it uses multiple isomorphic code fragments and the goto operator to manage state, reducing overhead. Theoretical analysis shows it outperforms standard multi-way merge methods.

Based on this, the author presents guided K-merge sort, a general-purpose sorting algorithm. Practical benchmarks indicate it can be up to 15% faster than conventional merge sort, depending on data type. The article covers the algorithm’s foundation, implementation, theoretical evaluation, and benchmarking results against existing merge-based sorts.