Helix-rc: An architecture-compiler co-design for automatic parallelization of irregular programs

Data dependences in sequential programs limit paralleliza-tion because extracted threads cannot run independently. Al-though thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to syn-chronize actual dependences counteract the benefits of paral-lelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing com-piler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a pro-cessor with 16 Intel Atom-like cores, show an average of 6.85× performance speedup for six SPEC CINT2000 benchmarks. 1