DYNAMIC E2E ANALYTIC TASK OPTIMIZATION, CODE GENERATION AND JIT COMPILE BASED ON LAYERED OPERATOR CATALOG AND DATA FLOW GRAPH REPRESENTATION OF ANALYTIC TASK

摘要

The objective of the present invention is to provide end-to-end analytic task optimization and dynamic code generation using a JIT compile based on the reuse of an operator, the securing of algorithm scalability, and data flow graph representation through the layered operator catalog structure of complex analysis algorithm. The present invention comprises a method for modeling a data flow graph by combining code templates which are stored after the code templates of a complex analysis algorithm are defined and the defined code templates are stored in a layered operator catalog; and a method for executing an end-to-end optimization of a data flow through a JIT compile.;COPYRIGHT KIPO 2017