Developments:CCC

From UNIDAPT.org

Home > Developments

Continuous Collective Compilation Framework simplifies and automates collection of heterogeneous optimization information from multiple users (architectural configurations, programs, datasets, run-time behavior, etc) to enable collective optimization for performance, code size, power, compilation time, etc. It supports both global program optimizations (compiler flags) and fine-grain optimizations on function, loop and instruction level through plugins if a compiler supports our Interactive Compilation Interface.

All data is stored in the Collective Optimization Database to enable further analysis and data processing using statistical techniques and machine learning. This framework is highly portable and have been tested on various platforms including x86, x8664, IA64, ARC, Longson/Godson, etc. It uses MILEPOST GCC to enable iterative compilation and program feature extraction. We ported multiple benchmarks such as EEMBC, SPEC2006, SPEC2000, SPEC95, MiBench with MiDataSets. We use PAPI library to collect dynamic program features.

We believe that using CCC Framework and collective optimization techniques will automate and simplify the process of developing and optimizing new architectures, compilers, operating systems and programming environments using machine learning, dynamic adaptation and bio-inspired techniques. We believe that these approaches are critical to overcome the complexity of computing systems and improve their performance, power consumption, system size and fault-tolerance automatically while reducing their cost and time to market. We hope that our techniques will boost innovation in science and industries (bioinformatics, medicine, physics, chemistry, finances, gaming, etc) that demand ever-increasing computing resources while placing strict requirements on systems.