Launching an Extreme-scale ACME Prototype - Transport (LEAP-T)
Substantial and ongoing improvements to coupled models of the Earth system have enabled a growth of confidence in predictions of future global climate change. Providing quantified predictions regarding the impacts of global climate change on regional scales and human systems, requires advanced global models that resolve multiple spatial scales, capture missing processes, and better quantify the uncertainties of model results. The U.S. Dept. of Energy (DOE) Accelerated Climate Modeling for Energy (ACME) project is chartered to develop an advanced Earth system modeling capability to address these needs and has targeted upcoming 100 Petaflop computing systems being delivered to DOE Leadership Class Facilities (LCFs) in 2018--2019. LEAP is preparing ACME for the next generation of advanced architectures and launching ACME on a trajectory toward extreme-scale simulation by delivering two capabilities. First, we are implementing a new transport algorithm that is optimal for the large tracer simulations planned by ACME and provides higher order accuracy. Second, we are optimizing key ACME transport "mini-apps," using new data layout strategies and data structures, as well as new approaches to task parallelism.
Project was started in September 2015 and will last two years.