Instruction Selection for Embedded DSPs with Complex Instruction
Chen Yan-haoa
Abstract
Chen Yan-haoa
Abstract
In this paper,we mainly explore the problem of code selection in code generation for embedded digital signal processor.We explore how such complex instructions affect the detection of optimal tree covering.Meanwhile,we introduce a two-phase model for code selection which can exploit available instruction parallel efficiently.Compared with previous work,this technology can significantly improve the code quality,which is demonstrated for a widespread DSPs.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper,we mainly explore the problem of code selection in code generation for embedded digital signal processor.We explore how such complex instructions affect the detection of optimal tree covering.Meanwhile,we introduce a two-phase model for code selection which can exploit available instruction parallel efficiently.Compared with previous work,this technology can significantly improve the code quality,which is demonstrated for a widespread DSPs.
Key concepts: Computer science, Code (set theory), Selection (genetic algorithm), Exploit, Digital signal processor, Parallel computing, Digital signal processing, Computer architecture