Dynamic Run›time Architecture Techniques for Enabling Continuous Optimization

Future computer systems will integrate tens of multithreaded processor cores on a single chip die, resulting in hundreds of concurrent program threads sharing system resources. These designs will be the cornerstone of improving through-put in high-performance computing and server environments. However, to date, appropriate systems software (operat-ing system, run-time system, and compiler) technologies for these emerging machines have not been adequately explored. Future processors will require sophisticated hardware mon-itoring units to continuously feed back resource utilization information to allow the operating system to make opti-mal thread co-scheduling decisions and also to software that continuously optimizes the program itself. Nevertheless, in order to continually and automatically adapt systems re-sources to program behaviors and application needs, specic run-time information must be collected to adequately enable dynamic code optimization and operating system schedul-ing. Generally, run-time optimization is limited by the time required to collect proles, the time required to perform op-timization, and the inherent benets of any optimization or decisions. Initial techniques for eectively utilizing run-time information for dynamic optimization and informed thread scheduling in future multithreaded architectures are presented