We propose to radically extend the frontiers of two major themes incomputing, parallelism and dynamism, and develop a novel paradigm ofcomputing: dynamic-parallelism. To this end, we will follow twolines of research. First, we will develop techniques for extractingefficiency and high performance from parallel programs written inhigh-level programming languages. Second, we will develop thedynamic-parallelism model, where computations can respond to a widevariety of dynamic changes to their data automatically andefficiently, by developing novel abstractions (calculi), high-levelprogramming-language constructs, and compilation techniques. Theresearch will culminate in a language that extends the C programminglanguage with support for parallel and dynamic-parallel programming.The proposal is motivated by urgent needs driven by the advent ofmulticore chips, which is making parallelism mainstream, and theincreasing ubiquity of software, which requires applications tooperate on highly dynamic data. These advances demand parallel andhighly dynamic software, which remains too difficult and laborintensive to develop. The urgency is further underlined by theincreasing data and problem sizes---online data growsexponentially, doubling every few years---that require similarlypowerful advances in performance.The proposal will achieve profound impact by dramatically simplifyingthe development of high-performing dynamic and dynamic-parallelsoftware. As a result, programmer productivity and software qualityincluding correctness, reliability, performance, and resource (e.g.,time and energy) consumption will improve significantly. The proposalwill not only open new research opportunities in parallel computing,programming languages, and compilers, but also in other fields whereparallel and dynamic problems abound, e.g., algorithms, computationalbiology, geometry, graphics, machine learning, and software systems.