Binary translation method combining static optimization

Abstract

The invention relates to a dynamic binary translation method combined with static optimization. The method comprises: inserting a detection instruction into a translated code during the first execution of a source program; collecting abundant subdivision information and realizing persistence of the subdivision information and translated object codes at the end of program execution; carrying out various optimizations of the translated object codes in a static state by means of stored subdivision information; and directly loading the optimized object codes during executions after that of the program. The novel translation mode provided by the invention can eliminate profiling generated during the operation of a dynamic binary translator, optimize overhead and reduce translation overhead; through execution of optimization in a static state, the dynamic binary translation method can make use of comprehensive subdivision information without taking optimization overhead into consideration and improves the quality of optimized codes, thereby obviously improving program performance during second and subsequent executions.

Description

technical field [0001] The invention relates to a dynamic binary translation method combined with static optimization, which is used for improving the execution performance of a program of a source machine platform on a translator. The invention belongs to the technical field of binary translation. Background technique [0002] Dynamic binary translation is used to dynamically translate the binary code corresponding to the source machine platform into the binary code that can run on the target machine platform, thus providing a way for software transplantation and system performance improvement. When the dynamic binary translator is running, it only translates the code fragments that are executed, which is completely transparent to the user. The translated code can be optimized by using the dissection information when the program is running, and these dynamic execution information cannot be collected statically. [0003] Profiling (dissection) refers to the process of coll...

AI Technical Summary

Problems solved by technology

[0009] 2. The optimization program itself will consume performance, and to obtain better optimization results, the optimization algorithm will generally be more complex, which will increase performance overhead

[0010] 3. It is impossible to obtain the behavior information of the entire running period of the program during the dynamic period. The subdivision information used by dynamic optimization is generally only the behavior information of the program running to a certain stage, which cannot reflect the whole picture of the program running behavior, which will lead to some dynamic optimization ( For example, the optimization effect of instruction branch jump prediction, etc.) is discounted

[0011] 4. If the same program needs to be run multiple times on the dynamic binary translator, the process of translating the binary code corresponding to the source machine platform into the binary code of the target machine platform will exist every time the program runs, and this part of the translation time will also bring overhead

Although this solution can improve the execution performance, its advantage is only to use the dissection information to generate better object code, not to optimize the object code or execution flow itself, and it cannot greatly improve the actual execution performance.