Top-Down Merge Sort Network for Parallel List Merging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sorting networks require multiple phases to merge multiple sorted input lists into a single sorted output list, which is inefficient and time-consuming.
Innovation Solution
The Top Down Merge Sort (TDMS) network efficiently merges sorted input lists by mapping values into a 2-D array and performing a sequence of sorting operations, reducing the number of phases required through unique mapping and sorting techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional sorting networks use multiple phases to merge multiple sorted input lists, then the merging process is reliable and complete, but the sorting time and number of phases increase significantly
Solution Approach 1:
The patent transforms the traditional one-dimensional sequential merging process into a two-dimensional parallel processing structure. Multiple sorted input lists are arranged in a 2-D array format where rows represent input lists and columns represent element positions. This dimensional transformation enables simultaneous comparison and exchange operations across multiple elements in parallel, reducing the number of phases from O(n) in traditional methods to O(log n) in the TDMS network while maintaining complete merging reliability.
Solution Approach 2:
The merging process is segmented into distinct phases: column sorting phase, row sorting phase, and diagonal sorting phase. Each phase handles specific comparison and exchange operations on particular segments of the 2-D array. This segmentation allows independent parallel execution of multiple sorting operations simultaneously, reducing overall sorting time while ensuring all elements are properly merged through systematic phase completion.
2Reliability
If traditional sorting networks use multistage 2-way merge processes, then the sorting operation is systematic and reliable, but the device complexity and number of sorting stages increase
Solution Approach 1:
The patent combines multiple 2-way merge operations into a single unified Top Down Merge Sort network structure. Instead of implementing separate multistage 2-way merge processes for each pair of input lists, the TDMS network integrates all merging operations into one systematic process using 2-D array organization. This combination reduces the number of sorting stages from multiple sequential 2-way merges to a condensed sequence of column, row, and diagonal sorting phases, decreasing device complexity while maintaining systematic reliability.
3Manufacturing precision
If existing sorting networks require many sorting phases to merge input lists, then complete sorting is achieved, but the productivity and data processing efficiency decrease
Solution Approach 1:
The TDMS network implements continuous useful action through its three-phase structure where column sorting, row sorting, and diagonal sorting phases proceed without interruption. Each phase continuously processes multiple elements in parallel rather than sequentially completing one element at a time. This continuous parallel processing maintains sorting accuracy through systematic comparison and exchange operations while dramatically improving data processing efficiency by eliminating idle time between operations.
Solution Approach 2:
The network performs preliminary organization of input lists into a 2-D array structure before the actual sorting phases begin. This preliminary action arranges elements in positions that facilitate efficient comparison and exchange during the subsequent column, row, and diagonal sorting phases. By pre-organizing the data structure, the system reduces the number of phases needed for complete sorting while maintaining accuracy, thereby improving overall productivity.
Data Source
AI summary
A method and apparatus for merging and sorting a plurality of lists. The method can be implemented in, and the apparatus can comprise, hardware and all phase sorting operations are preferably performed in parallel. Thus, a multistage sorting process can be implemented with 1 or more 2-sorters operating in parallel in each stage. This enables two or more lists to be sorted and merged very quickly. The input lists are preferably setup in an input array before the merge sorting operation is performed.