Machine learning and Optimization-guided Compilers for Heterogeneous Architectures (MOCHA)

Machine Learning and Optimization-Guided Compilers for Heterogeneous Architectures (MOCHA) seeks to build a new generation of compiler technology to realize the full potential performance of heterogenous architectures. MOCHA will develop data-driven methods, Machine Learning, and advanced optimization techniques to rapidly adapt to new hardware components with little human effort and facilitate optimal allocation of computation to heterogeneous components.

All responsible sources capable of satisfying the Government's needs may submit a proposal that shall be considered by DARPA. See the Eligibility Information section of the BAA for more information.

BAA Coordinator MOCHA@darpa.mil