PARALLEL CONSTRUCTION FOR GRAPH MODEL OF LONGEST COMMON NON-SUPERSTRING USING COMPUTE UNIFIED DEVICE ARCHITECTURE

摘要

PURPOSE: A parallel production method of a longest common non-superstring graph model using computer unified device architecture is provided to use a shared memory, thereby increasing performance. CONSTITUTION: A prefix graph for a string set is generated by using compute unified device architecture. Strings in the string set are arranged to a parallel radix sort. A peak set of the prefix graph is generated corresponding to the parallel radix sort. The strings are divided into a block for a character type of n-th location by arranging the strings. [Reference numerals] (AA) Arrangement; (BB) Arrangement; (CC) String set; (DD) Step 1; (EE) Step 2