Project Description

The unveiling of the large-scale structure of the Universe by modern sky surveys has transformed cosmology, fundamental physics, and astrophysics. The resulting cosmological Standard Model is a remarkable success, yet it raises deep questions about the nature and origin of its fundamental constituents and assumptions, e.g., dark energy, dark matter, and primordial fluctuations. Next-generation observatories unprecedented in their scale, cost, capabilities and complexity, are targeted at answering these questions, and making discoveries beyond the Standard Model. But their promise can be realized only with a theory, modeling, and simulation effort fully as revolutionary as the surveys themselves. Multi-wavelength simulations play a crucial role in interpreting the observations and providing predictions for different cosmological models and scenarios at an exquisite level of detail. In this project we will build upon some already available simulations and carry out new simulations of the Standard Model and beyond. The simulations and resulting sky maps will be used to support the analysis of ongoing DOE-funded cosmological surveys, to prepare for upcoming surveys, and to plan future observational campaigns. We will carry out a set of 29 simulations to finalize the creation of the Mira-Titan Universe suite covering more than 110 cosmological models. These state-of-the-art simulations span a range of cosmologies, including models with a dynamical dark energy equation of state and massive neutrinos. Each simulation covers a volume of (2100Mpc)3 and evolves close to 33 billion particles. This set of simulations will enable us to explore subtle differences in cosmological observations due to variations in cosmological parameters. We will be able to investigate and develop new cosmological probes and cross-correlations that provide information about the best possible strategies to extract signals from non-CDM cosmologies and break possible parameter degeneracies. These simulations will be important for ongoing and upcoming surveys alike. We will use this ALCC allocation, to 1) carry out the last remaining cosmological models to create the complete Mira-Titan Universe simulation set, 2) develop a range of new emulators (prediction tools) and improve already existing ones by adding the new cosmological models, 3) transform the simulations into synthetic skies that closely mimic actual observational data, 4) curate the data in such a way that the results can be made accessible to the community. We will take advantage of two currently available architectures, BG/Q (Mira) and GPUaccelerated systems (Titan). Our Hardware/Hybrid Accelerated Cosmology Code (HACC) framework has been shown to deliver excellent performance on these machines and in combination with our extensive analysis toolkit, CosmoTools, we are in an optimal position to not only carry out these simulations but also to extract survey relevant science from them during this ALCC project. We will build multi-wavelength synthetic maps from these simulations, including modeling of the kinematic and thermal Sunyaev-Zel’dovich effect, weak lensing maps, and optical catalogs. We have recently started to develop a data portal for HACC simulation data. Our data hub utilizes Petrel, a data management and sharing project, located at the Argonne Leadership Computing Facility. Petrel offers fast data transfer mechanisms and authentication via Globus, enabling very convenient access to the data. We will use the data hub for providing access to the simulation products from this ALCC to the wider cosmology community.

Allocation History

Source Hours Start Date End Date