TransCache management method based on hot degree of code in dynamic binary translation

Abstract

The present invention provides a translation cache management method based on code popularity in dynamic binary translation. The method includes the steps of: dividing the entire translation cache into two levels of cache; when there is a new basic block to be translated, according to the basic block Find the code block in the hot code cache according to the address; look up the code block in the sub-code cache according to the address of the basic block; calculate the space required for caching the code block; clear the sub-code cache; translate the basic block, and store the translated code block into the sub-code cache; execute the code block; count the number of executions of the code block, if the number of executions is greater than the threshold, the code block is cached into The hot code cache; adding the basic block into the slow hash table, and adding the result into the fast hash table.

Description

technical field [0001] The invention relates to a dynamic binary translation cache management technology in the field of computer applications, in particular to a translation cache (TransCache) management method based on code heat in dynamic binary translation. Background technique [0002] In recent years, binary translation technology has become an important means to implement code transplantation by software. Since the research on binary translation and related optimization technology is of great significance in code porting, virtual machine, network application, computer security and hardware development, etc., it has become one of the hotspots of modern computer research. [0003] Dynamic binary translation technology is a main direction for the future development of binary translation technology. It dynamically translates the binary program of the source target machine into the binary program of the target machine. The basic working principle of dynamic binary transl...

AI Technical Summary

Problems solved by technology

The basis for the implementation of the full clear strategy based on the working set is that the execution of the program requires a clear phase distinction. The non-replacement strategy requires the TransCache space to be large enough. The largest block priority replacement strategy only considers the spatial locality of the program, but does not consider the temporal locality of the program. performance characteristics

Images