Optimizing assembled code for execution using execution statistics collection, without inserting instructions in the code and reorganizing the code based on the statistics collected

摘要

In one aspect, a software development technique is capable of efficiently organizing for execution a conditional code segment having multiple associated conditional paths. The development technique employs in association with each path of the code segment, a probability compiler directive which dictates to the compiler a probability for satisfying a conditional test of the associated path. In another aspect, a system/process is capable of optimizing organization of assembled program code for a code's particular execution environment. This optimization approach tunes assembled code organization for each specific execution environment, employing run-time based statistical data collected during performance execution of the assembled code. The execution environment, consisting of hardware, software, and other factors such as workload, input data, etc., can also be collected and employed by an optimizer unit to best reorganize the assembled program code for a current execution environment.