Aiichiro Nakano - US grants

This project will develop a scalable parallel computing framework for high-end computational research, which will achieve scalability beyond tightly coupled Teraflop architectures, i.e., for distributed supercomputing on multiple Teraflop computers as well as on future Petaflop computers with deep memory hierarchy. To accomplish this goal, the PI will conduct the following research tasks:

Topology-preserving computational-space decomposition to minimize the number of messages using a structured message-passing scheme;
Wavelet-based adaptive load balancing in dynamic, heterogeneous metacomputing environment using simulated annealing to minimize load imbalance and message sizes;
Recursive and reconfigurable grouping of processors with message renormalization and computation/communication overlapping to hide latency at each grouping level;
Spacefilling-curve-based adaptive data compression --- in situ processing of interoperable compressed data to reduce message sizes with user-controlled error bound.

A suite of scalable scientific programs developed within the new framework will be disseminated as a computational-scientist's toolkit through a Web portal to have significant impacts on high-end computational research, including the design of quantum-dot architectures for future quantum computing.

This award was made on a collaborative proposal submitted to the Division of Materials Research under the Information Technology Research solicitation NSF-04-012. The Division of Materials Research, the Chemistry Division, and the Division of Computing and Communications Foundations fund this award. The other proposals in this multidisciplinary collaborative are 0427177 and 0427540 and involve investigators at Caltech and Purdue. Research activities covered by this award fall under the National Priority Area, "Advances in Science and Engineering," and the Technical Focus Area, "Innovation in Computational Modeling or Simulation in Research." This award supports computational research and algorithm development with the aim of developing new modeling tools for materials failure and with the further aim of applying these tools to advance the understanding of stress corrosion cracking. This award also supports related educational activities some of which involve underrepresented groups.

The PIs aim to develop a scalable parallel and distributed computational framework consisting of methods, algorithms, and integrated data handling and visualization tools for: 1) an accurate quantum mechanical-level (QM) description; 2) reactive force fields (ReaxFF) to describe chemical reactions and polarization; 3) molecular dynamics (MD) simulations to extract atomistic mechanisms of SCC; 4) accelerated dynamics for long-time behavior to obtain parameters directly comparable to experiments; and 5) "atomistically informed" continuum models to reach macroscopic length and time scales. Automated model transitioning by novel techniques will be employed to embed higher fidelity simulations inside coarser simulations only when and where they are required, while controlled error propagation will ensure the overall accuracy of the results. The PIs plan to use this hierarchical multiscale computational framework to study stress corrosion cracking (SCC) of aluminum, iron, and nickel-aluminum superalloys in gaseous and aqueous environments. These materials are used widely in industrial applications and their performance and lifetime are often severely limited by stress corrosion in environments containing oxygen and water. Simulations will be used to extract an atomic-level understanding of the basic mechanisms underlying SCC. The PIs plan to investigate SCC inhibition by ceramic coatings (e.g., alumina and silicon carbide), self-assembled monolayers (e.g., oleic imidazolines), and by microorganisms (e.g., Shewanella oneidensis strain MR-1).

The PIs will deliver software tools having broad applicability across scientific disciplines and industry. This award supports annual computational science workshops for undergraduate students and faculty mentors from underrepresented groups. Workshops will be organized to foster close interactions between underrepresented minority graduate students at US institutions and postdoctoral level counterparts from Latin American institutions. Undergraduate students will be involved in the research through summer research experiences; at least half are expected to be from underrepresented groups. The PIs will also assist minority institutions in developing computational science curricula, and mentor early-career faculty from minority institutions and EPSCoR states.

This award also supports education. Elements of the PIs' education program include: 1) a unique graduate course jointly taught by USC and Caltech faculty emphasizing hands-on experience in hierarchical multiscale material simulations; 2) a dual-degree program at USC offering graduate students the opportunity to obtain a PhD in the physical sciences or engineering and an MS in computer science with specialization in high performance computing and simulations; and 3) summer research experiences for undergraduate students involving a total immersion course in computational science followed by research in simulation, parallel algorithms and visualization.
%%%
This award was made on a collaborative proposal submitted to the Division of Materials Research under the Information Technology Research solicitation NSF-04-012. The Division of Materials Research, the Chemistry Division, and the Division of Computing and Communications Foundations fund this award. The other proposals in this multidisciplinary collaborative are 0427177 and 0427540 and involve investigators at Caltech and Purdue. Research activities covered by this award fall under the National Priority Area, "Advances in Science and Engineering," and the Technical Focus Area, "Innovation in Computational Modeling or Simulation in Research." This award supports computational research and algorithm development with the aim of developing new modeling tools for materials failure and with the further aim of applying these tools to advance the understanding of stress corrosion cracking. This award also supports related educational activities some of which involve underrepresented groups.

Stress corrosion cracking (SCC) is a complex technological and economic problem involving premature and catastrophic failure of materials due to an insidious combination of mechanical stresses and chemically aggressive environments. Safe and reliable operation of structural systems are endangered by uncertainties in SCC, the reduction of which could have enormous economic impact. The PIs plan to develop computational tools that contain essential physics across a wide range of length and time scales to achieve an atomic-level mechanistic understanding of SCC. Because of the large number of atoms and complex physical and chemical processes, these tools will be able to manage distributed computing resources and focus them on SCC simulation.

The PIs plan to use these tools to study SCC of aluminum, iron, and nickel-aluminum superalloys in gaseous and aqueous environments. These materials are used widely in industrial applications and their performance and lifetime are often severely limited by stress corrosion in environments containing oxygen and water. Simulations will be used to understand the basic mechanisms underlying SCC. The PIs plan to investigate how various coatings and microorganisms inhibit SSC.

This award also supports education. Elements of the PIs' education program include: 1) a graduate course jointly taught by USC and Caltech faculty emphasizing hands-on experience in hierarchical multiscale material simulations; 2) a dual-degree program at USC offering graduate students the opportunity to obtain a PhD in the physical sciences or engineering and an MS in computer science with specialization in high performance computing and simulations; and 3) summer research experiences for undergraduate students.

The PIs will deliver software tools having broad applicability across scientific disciplines and industry. This award supports annual computational science workshops for undergraduate students and faculty mentors from underrepresented groups. Workshops will be organized to foster close interactions between underrepresented minority graduate students at US institutions and postdoctoral level counterparts from Latin American institutions. Undergraduate students from underrepresented groups will be involved in the research. In addition, the PIs will assist minority institutions in developing computational science curricula and mentor early-career faculty from minority institutions and EPSCoR states.
***

This project systematically addresses the enormous complexity of mapping applications to current and future parallel platforms - both scalable parallel architectures consisting of tens of thousands of processors and distributed systems comprised of collections of these and other resources. By integrating the system layers - domain-specific environment, application program, compiler, run-time environment, performance models and simulation, and workflow manager -- and through a systematic strategy for application mapping, the project will exploit the vast machine resources available in such parallel platforms to dramatically increase the productivity of application programmers.

The key contribution of the project will be a systematic solution for performance optimization and adaptive application mapping -- a large step towards automating a process that is currently performed in an ad hoc way by programmers and compilers -- so that it is feasible to obtain scalable performance on parallel and distributed systems consisting of tens of thousands of processing nodes. The application components will be viewed as dynamically adaptive algorithms for which there exist a set of variants and parameters that can be chosen to develop an optimized implementation. Knowledge representation and machine learning techniques utilize this domain knowledge and past experience to navigate the search space efficiently.

CSR-AES: Intelligent Optimization of Parallel and Distributed Applications

ABSTRACT
This project derives a systematic solution for performance optimization and adaptive application mapping to obtain scalable performance on parallel and distributed systems consisting of tens of thousands of processing nodes. With expert domain scientists in molecular dynamics (MD) simulation, we expect to achieve performance levels on MD codes even better than what has been derived manually after years of development and many ports to a variety of architectures.
The application components are viewed as dynamically adaptive algorithms for which there exist a set of variants and parameters that can be searched to develop an optimized implementation. A workflow is an instance of the application where nodes represent application components and dependences between the nodes represent execution ordering constraints. By encoding an application in this way, we capture a large set of possible application mappings with a very compact representation. The system layers explore the large space of possible implementations to derive the most appropriate solution. Because the space of mappings is prohibitively large, the system captures and utilizes domain knowledge from the domain scientists and designers of the compiler, run-time and performance models to prune most of the possible implementations. Knowledge representation and machine learning utilize this domain knowledge and past experience to navigate the search space efficiently.
This multidisciplinary approach impacts the state-of-the-art in the sub-fields of compilers, run-time systems, machine learning, knowledge representation, and accelerates advances in MD simulation with far more productive software development and porting. More broadly, this research enables systematic performance optimization in other sciences.

TECHNICAL SUMMARY:

This award is made on a proposal submitted to the PetaApps Solicitation. The Office of Cyberinfrastructure, the Division of Materials Research and Office of Multidisciplinary activities in the Mathematical and Physical Sciences Directorate, the Engineering Directorate, and the Computer and Information Science and Engineering Directorate contribute funds to this award.

This PetaApps project focuses on hybrid quantum mechanical-atomistic-mesoscale simulations of ion transport and translocation of biopolymers such as DNA and RNA through nanometer scale pores and channels in silica and silicon nitride membranes. The PIs aim to develop a predictive hierarchical petascale simulation framework for: (1) Highly accurate quantum mechanical simulations to describe chemical processes in translocating biopolymers; (2) multibillion-atom molecular dynamics simulations for structural properties and dynamical processes of biopolymers in confined fluidic environments in solid state membranes, with interatomic interactions validated by quantum mechanical calculations and key experiments; (3) hybrid molecular dynamics and adaptive lattice Boltzmann simulations in which molecular dynamics is embedded close to the surfaces of nanopores/nanochannels and lattice Boltzmann in the rest of the fluid; (4) accelerated dynamics approaches to reach macroscopic time scales for direct comparison with experimental data; (5) meta-scalable, self-tuning multicore parallel simulation algorithms; and (6) automated model transitioning to embed higher fidelity simulations inside coarser simulations on demand with controlled error propagation to quantify uncertainty.

After validation, this hierarchical petascale simulation framework will be used to study: (1) Translocation kinetics and dynamics of DNA through silica and silicon nitride nanopores; (2) electronic properties of translocating DNAs for sequential identification of nucleotides; (3) ionic screening of surface charges in nanopores/nanochannels; (4) streaming electrical current generated by pressure-driven liquid flow in individual silica nanochannels as a function of channel height, pressure gradient, and salt concentration; (5) pressure-driven DNA transport in confined silica channels for novel diagnostic applications such as artificial gels and entropic trap arrays; and (6) surface functionalization, polarity switching, and transient response of silica nanotube, nanofluidic transistors.

This project supports training a new generation of graduate students to develop the tools needed to attack complex system level problems. They will learn to combine theory, modeling, and high performance computer simulation. Students will participate in a dual-degree program in which they will fulfill Ph.D. requirements within their own discipline and master?s degree requirements in computer science with specialization in high performance computing and simulations.
This award also supports the computational science workshops for underrepresented groups. Undergraduate students and faculty mentors from Historically Black Colleges and Universities and Minority Serving Institutions participate in a special one-week intense hands-on experience in parallel computing and immersive and interactive visualization. African American, Hispanic and Native American students will be recruited through USC?s Center for Engineering Diversity and women through USC?s Women in Science and Engineering Program.

NON-TECHNICAL SUMMARY:

This award is made on a proposal submitted to the PetaApps Solicitation. The Office of Cyberinfrastructure, the Mathematical and Physical Sciences Directorate, the Engineering Directorate, and the Computer and Information Science and Engineering Directorate contribute funds to this award.

This award supports the development of software for the most advanced, ?petascale,? high performance supercomputers that will enable simulations that can capture phenomena that span across a range of length and time scales. The PIs will focus on a problem of particular importance, how biomolecules move through nanometer-sized pores in inorganic materials like silica and silicon nitride. The simulation can capture detailed physics of the problem and may illuminate possible applications to sequencing DNA and RNA molecules. The PIs will also focus on how charged atoms and molecules move through channels with dimensions on nanometer length scales more generally. There are potential applications to evolving ?lab-on-a-chip? technologies that seek to miniaturize laboratory analysis functions to the size of electronic device chips.

Developed software will be distributed and can be used by a broad community of researchers in a variety of disciplinary and multidiscplinary research involving materials research, chemistry, engineering, physics, and nanotechnology.

This project supports training a new generation of graduate students to develop the tools needed to attack complex system level problems. They will learn to combine theory, modeling, and high performance computer simulation to solve complex problems. Students will participate in a dual-degree program in which they will fulfill Ph.D. requirements within their own discipline and master?s degree requirements in computer science with specialization in high performance computing and simulations.

This award also supports the computational science workshops for underrepresented groups. Undergraduate students and faculty mentors from Historically Black Colleges and Universities and Minority Serving Institutions participate in a special one-week intense hands-on experience in parallel computing and immersive and interactive visualization.

PETASCALE SIMULATIONS OF DNA DYNAMICS AND SELF-ASSEMBLY
Priya Vashishta?PI, Rajiv K. Kalia, Aiichiro Nakano (University of Southern California)
Ananth Grama (Purdue University)

DNA translocation through solid-state nanopores and nanofluidic channels underlie ?lab-on-a-chip? technology and solid-state nanopore ?microscopy? for molecular structure and high-speed sequencing. Highly efficient methods for directed self-assembly of DNA offer unprecedented opportunities for the synthesis of novel genes, chromosome mapping, biosensors, molecular machines, nanoelectronics and nanomechanical systems, and formulations of mesoscopic structural motifs as building blocks of emerging periodic and aperiodic nanostructures consisting of DNAs.
This project involves the study of DNA self-assembly and translocation through nanometer-scale pores in silica and silicon nitride membranes using a predictive hierarchical petascale simulation framework consisting of: (1) Highly accurate quantum mechanical (QM) simulations to describe chemical processes in DNA translocation and concatenation; (2) multibillion-atom molecular dynamics (MD) simulations for structural properties and dynamical processes of DNAs in confined fluidic environments, with interatomic interactions validated by QM calculations and key experiments; (3) hybrid MD and adaptive lattice Boltzmann (LB) simulations in which MD is embedded in translocation/concatenation regions, and LB in the rest of the fluid; (4) accelerated dynamics approaches to reach macroscopic time scales for direct comparison with experimental data; (5) metascalable, self-tuning, multicore parallel simulation algorithms; and (6) automated model transitioning to embed higher fidelity simulations inside coarser simulations on demand with controlled error propagation. A metascalable (or ?design once, scale on new architectures?) parallel application-development framework is also being developed for first-principles simulations of directed DNA self-assembly.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

This project establishes a dedicated computing platform for microsecond simulations to study DNA self-assembly and translocation through solid-state nanopores. The project uses a predictive hierarchical petascale simulation framework to study: Translocation kinetics and dynamics of DNAs through solid state nanopores; electronic properties of translocating DNAs for sequencing nucleotides; ionic screening of surface charges in nanopores; pressure-driven DNA transport in confined silica channels; and shear-induced DNA self-assembly.

The computing platform will also support computer science research in techniques for the parallelization of such simulations, and for the integration of multi-scale, multi-phenomena simulation codes for molecular biology and biological materials science.

Petascale simulations of DNA translocation through solid-state nanopores and nanofluidic channels underlie "lab-on-a-chip" technology and solid-state nanopore "microscopy" for molecular structure and high-speed sequencing.

The infrastructure will help in training a new generation of graduate students. Students participate in a dual-degree program in which they do a PhD in physical sciences or engineering and a master's degree in computer science. The infrastructure also strengthens the annual computational science workshops for underrepresented groups, in which undergraduate students and faculty mentors from Historically Black Colleges and Universities and Minority Serving Institutions acquire hands-on experience in parallel computing.

Further information on the project can be found at the project web page: http://cacs.usc.edu/cri/index.php

The discovery of RNA interference mediated gene silencing by double-stranded RNA is one of the most exciting recent developments in the biological and biomedical sciences. RNAi utilizes a pathway whereby a double-stranded small interfering RNA (siRNA) targets and destroys complementary mRNA in eukaryotic cells to regulate gene expression. Since its discovery just over a decade ago scientists have investigated the potential use of siRNA-based treatment of many diseases such as cancer, liver cirrhosis, hepatitis B, human papillomavirus, and hypercholesterolemia. In principle, siRNA has much broader therapeutic applications than other types of drugs because siRNA can be synthetically designed to silence any gene via the RNAi machinery. Therefore, the primary challenge for therapeutic applications is efficient and non-toxic delivery of siRNA to cells within tissues of interest. The research goal of this CDI project is to design efficient delivery systems for siRNA via supercomputer simulations. Multimillion-atom simulations are performed to study: (1) the effect of siRNAs on the molecular structure of lipid membranes and how structural changes affect the membrane permeability; (2) molecular mechanisms by which siRNA molecules attached to gold nanoparticles cross the membrane and go into a cell; and (3) delivery of siRNAs encapsulated in liposomes by ultrasound. Computational technologies will be used to enable automated design of efficient siRNA delivery systems. This project is promoting scholastic and professional excellence among students. A dual-degree program has been established which gives students the opportunity to obtain a Ph.D. in the physical sciences or an engineering discipline and an M.S. in computer science with specialization in high performance computing and simulations. This research team will also organize computational science workshops for undergraduate students and faculty mentors from Historically Black Colleges and Universities and Minority Serving Institutions.

Large scale scientific simulations are used in a range of application domains including materials science, climate modeling, combustion and others. These simulations are often limited by the hardware on which they run, including the capacity of computational platforms and storage systems. The goals of these simulations may include finding rare but interesting events in simulation output, discovering common sequences of events, and discovering causality among events. Often, these scientific simulations consume all available computational resources on a high performance computing platform during a simulation run, and be forced to only sample data techniques to decrease the size of the simulation so as to make it possible to store, transfer and post-process the output data. Such data sampling reduces the quality of science results, since not all available data are utilized during analysis. This project aims to greatly improve the scale and quality of scientific simulation results using innovative "in situ" algorithms and machine learning techniques for rare event detection. This research will be validated using a large-scale materials science simulation, that of self-healing nanomaterial system capable of sensing and repairing damage in harsh chemical environments and in high temperature/high pressure operating conditions. Self-healing is of significance since it can improve the reliability and lifetime of materials while reducing the cost of manufacturing, monitoring and maintenance of high-temperature turbines, wind, solar energy and lighting systems. The research can be generalized to a range of scientific simulation domains that share the common goals of discovering rare and interesting events, sequences of events and causality among events. Finally, the research concepts and results will be incorporated into graduate level courses taught by the research team.

The goal of the project is to demonstrate the feasibility, performance and scalability of the research approaches in greatly improving the quality of exascale scientific simulations using in situ machine learning algorithms within a well-defined, reusable in situ software framework. The scope of the project includes: selecting a simplified, but representative, long-time material process suitable for super-state parallel replica dynamics (SPRD); developing in situ machine learning algorithms for rare event detection of super-state transitions; and studying library-based approaches to support the high performance coupling of exascale simulations with in situ machine learning algorithms. To accomplish the project goals, the following three objectives are defined: 1) Prove the feasibility, performance and scalability of in situ SPRD simulation for predicting long-time material processes; 2) Prove the feasibility, performance and scalability of in situ machine learning algorithms for rare-event detection of super-state transitions; and 3) Prove the feasibility, performance and scalability of in situ library-based approaches to coupling exascale simulations and machine learning algorithms.

The computing landscape is evolving rapidly. Exascale computers can perform unprecedented mathematical operations per second, while quantum computers have surpassed the computing power of the fastest supercomputers. Concomitantly, artificial intelligence (AI) is transforming every aspect of science and engineering. To address these rapid changes and challenges, this project will train a new generation of materials cyberworkforce, who will solve challenging materials genome problems through innovative use of advanced cyberinfrastructure (CI) at the exa-quantum/AI nexus. Further, the project will foster the adoption of exa-quantum/AI nexus technologies by a broad research community and beyond through a unique dual-degree PhD/MS program, undergraduate research to close the research-education gap, and broadening participation of women and underrepresented groups.

This project will develop training modules for a new generation quantum materials simulator named AIQ-XMaS (AI and quantum-computing enabled exascale materials simulator), which integrates exa-scalable quantum, reactive and neural-network molecular dynamics simulations with unique AI and quantum-computing capabilities to study a wide range of materials and devices of high societal impact such as optoelectronics and pandemic preparedness. CyberMAGICS (cyber training on materials genome innovation for computational software) portal will be developed as a single-entry access point to all training modules that include step-by-step instructions in Jupyter notebooks and associated tutorial slides/videos, while providing online cloud service for those who do not have access to computing platform. The modules will be incorporated into the open-source AIQ-XMaS software suite as tutorial examples, and they will be piloted in classroom and workshop settings to directly train 1,200 CI users at the University of Southern California (USC) and Howard University, with a strong focus on underrepresented groups. Broader reach and training will be accomplished through the portal and nanoHUB. Students trained in the dual-degree program will earn a PhD in materials science or physics; they will also earn either an MS in computer science specialized in high-performance computing and simulations, MS in quantum information science, or MS in materials engineering with machine learning. Undergraduate students will be mentored and trained by academic scholars in multidisciplinary fields as well as by scientists at national labs and industry. The project will further broaden participation through USC’s Women in Science and Engineering (WiSE) program and undergraduate research by underrepresented groups jointly supervised by USC and Howard faculty.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.