Transform of single line routine code to conjecture preexecute starting code

Abstract

In one embodiment a thread management method identifies in a main program a set of instructions that can be dynamically activated as speculative precomputation threads. A wait / sleep operation is performed on the speculative precomputation threads between thread creation and activation, and progress of non-speculative threads is gauged through monitoring a set of global variables, allowing the speculative precomputation threads to determine its relative progress with respect to non-speculative threads.

Description

technical field [0001] This invention relates to computing system software, and more particularly to thread management. Background technique [0002] Efficient operation of modern computing systems often requires support for multiple instruction "threads," each thread being a stream of instructions that provides a clear flow of control within a program. To increase overall system speed and responsiveness, multithreading can be implemented simultaneously by computing systems having multiple processors, where each processor supports a single thread. In more advanced computing systems, multithreading may be supported using processors having a multithreaded processor architecture capable of executing multiple threads simultaneously. Alternatively, a single processor can be multiplexed among threads after a fixed period of time, in a technique commonly referred to as time-sliced ​​multithreading. In another approach known as event-switched multithreading, a single processor swi...

AI Technical Summary

Problems solved by technology

However, this technique does not always work due to the synchronous release of threads

Images