Research & Development: Projects

Legacy Projects

Listed here are selected past research projects and topics in Computing that are no longer actively updated. These projects’ web pages do not support accessibility features.

AutomaDeD is a tool that automatically diagnoses performance and correctness faults in MPI applications. It has two major functionalities: identifying abnormal MPI tasks and code regions and finding the least-progressed task. The tool produces a ranking of MPI processes by their abnormality degree and specifies the regions of code where faults are first manifested.

Focus Area: 
Debugging and Correctness Tools, Performance Analysis Tools

About Babel

Babel is a high-performance language interoperability tool. The project is mainly developed at the Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory (LLNL). Babel started as an internal Lab Directed Research and Development (LDRD) project in 2000 and has been under constant development since then. It is now funded mainly under the U.S. Department of Energy (DOE) Office of Science’s SciDAC program.

Babel currently fully supports C, C++, FORTRAN 77, Fortran 90/95, Fortran 2003/2008, Python, and Java. It won a prestigious R&D 100 award in 2006 for “The world’s most rapid communication among many programming languages in a single application.

Our tool is based on the Scientific Interface Description Language (SIDL). SIDL is tailored to the needs of the scientific community and features support for complex numbers, structs, and dynamic multidimensional arrays. Babel is a compiler that generates glue code from SIDL interface descriptions. SIDL provides a modern object oriented programming model, even on top of traditional procedural languages. This includes automatic reference counting and resource (de)allocation. Code written in one language can be called from any of the supported languages.

Babel focuses on high-performance language interoperability within a single address space. Full support for Remote Method Invocation (RMI) allows for the development of parallel distributed applications.

We also provide tools to automatically generate documentation from SIDL descriptions. An equivalent XML representation eases development of third-party tools.

Focus Area: 
Programming Models and Languages, Compiler Technology

Theg Extreme Resilient Discretization project (ExReDi) was established to address these challenges for algorithms common for fluid and plasma simulations.

Focus Area: 
Plasma Physics, High-order Finite Volume Methods, Multigrid and Multilevel Solvers

LLNL researchers are developing a truly scalable first-principles molecular dynamics algorithm with O(N) complexity and controllable accuracy, capable of simulating systems of sizes that were previously impossible with this degree of accuracy. By avoiding global communications, a practical computational scheme capable of extreme scalability has been implemented.

Focus Area: 
Materials Science/Molecular Dynamics

The code GEFIE-QUAD (gratings electric field integral equation on quadrilateral grids) is a first-principles simulation method to model the interaction of laser light with diffraction gratings, and to determine how grating imperfections can affect the performance of the compressor in a CPA laser system. GEFIE-QUAD gives scientists a powerful simulation tool to predict the performance of a realistic laser compressor.

Focus Area: 
Computational Sciences/Simulation, Ultra-short Pulse Laser Propagation

In response to a Department of Energy grid optimization challenge, the LLNL-led gollnlp team is developing the mathematical, computational, and software components needed to solve problems of the real-world power grid.

Focus Area: 
Computational Sciences/Simulation, Power Grid, Data Analytics and Management

To overcome the shortcomings of the analytical and architectural approaches to performance modeling and evaluation, we are developing techniques that emulate the behavior of anticipated future architectures on current machines. We are implementing our emulation approaches in what we call the GREMLIN framework. Using GREMLIN, we can emulate a combined effect of power limitations and reduced memory bandwidth and then measure the impact of the GREMLIN modifications.

Focus Area: 
Power-Aware Computing, Performance Analysis Tools

Livermore researchers are enhancing HARVEY, an open-source parallel fluid dynamics application designed to model blood flow in patient-specific geometries. Researchers will use HARVEY to achieve a better understanding of vascular diseases as well as cancer cell movement through the bloodstream. Establishment of a robust research platform could have direct impact on patient care. HARVEY is also an enabling capability for the BAASiC initiative.

Focus Area: 
Computational Biology

hzip 1.0.1 is a C++ library for lossless compression of structured and unstructured meshes composed of cells with hypercube topology.

Focus Area: 
Compression Techniques

As supercomputing power increases, the cost of communicating data both on- and off-node has become a critical factor affecting the overall performance of a parallel application. Livermore’s interconnection networks projects improve the communication and overall performance of parallel applications using interconnect topology-aware task mapping.

Focus Area: 
Interconnection Networks

The sheer size of data poses significant problems in all stages of the visualization pipeline, from offline pre-processing of simulation data, to interactive queries, to real-time rendering. Moreover, visualization data is often unstructured in nature, which further complicates its management and representation. The goal of this project is to develop techniques for reducing bandwidth requirements for large unstructured data, both explicitly, by making use of data compression, and implicitly, by optimizing the layout of the data for better locality and cache reuse.

Focus Area: 
Visualization and Scientific Data Analysis

MPI_T is an interface for tools introduced in the 3.0 version of MPI. The interface provides mechanisms for tools to access and set performance and control variables that are exposed by an MPI implementation. The latest versions of major MPI implementations are already providing MPI_T functionality, making it widely accessible to users. We have developed a set of MPI_T tools, Gyan and VarList, to help tool writers with the new interface.

Focus Area: 
Middleware for Parallel Performance Tools, MPI

NEWSO5 is a Fortran MPI performance code developed at LLNL. This code measures the performance of MPI nonblocking communications using the communication pattern relevant to two-dimensional stencil-based computations. The name of the code is derived from the common name for this communication pattern, NEWS, for the North-East-West-South directions used to determine communication endpoints. The code tests all possible grid layouts for the total number of MPI tasks in the job. The original code exchanges messages of a size determined at compile time. The code was modified as part of the PSE/ASDE project to allow multiple message sizes to be tested during a single run. The last version, which uses a harness coded in C to invoke the original Fortran main program, is available for download. The simplicity of NEWS05 combines with its high message volume to serve as good communication system diagnostic program, particularly with extended runs lasting more than 30 minutes.

Focus Area: 
MPI

ODEPACK is a collection of Fortran solvers for the initial value problem for ordinary differential equation systems. The collection is suitable for both stiff and nonstiff systems and includes solvers for systems given in both explicit and linearly implicit forms. The solvers are written in standard Fortran 77, with a few exceptions and with minimal machine dependencies.

ParFlow is a parallel, three-dimensional, variably saturated groundwater flow code that is especially suitable for large scale, high resolution problems ParFlow’s development and application has been on-going for more than 10 years and resulted in some of the most advanced numerical solvers and multigrid preconditioners for massively parallel computer environments that are available today.

Focus Area: 
Computational Fluid Dynamics

Livermore researchers have developed an algorithm for the numerical solution of a phase-field model of microstructure evolution in polycrystalline materials. The system of equations includes a local order parameter, a quaternion representation of local orientation, and species composition. The approach is based on a finite volume discretization and an implicit time-stepping algorithm. Recent developments have been focused on modeling solidification in binary alloys, coupled with CALPHAD methodology.

Focus Area: 
Materials Science/Molecular Dynamics

PMPI is a success story for HPC Tools, but it has a number of shortcomings. LLNL researchers aimed to virtualize the PMPI interface, enable dynamic linking of multiple PMPI tools, create extensions for modularity, reuse existing binary PMPI tools, and allow dynamic tool chain selection. The result is PnMPI, a thin, low-overhead wrapper library that is automatically generated from mpi.h file and that can be linked by default.

Focus Area: 
Middleware for Parallel Performance Tools

LLNL’s version of Qbox, a first-principles molecular dynamics code, will let researchers accurately calculate bigger systems on supercomputers.

Focus Area: 
Materials Science/Molecular Dynamics

LLNL researchers are developing a new algorithm for use with first-principles molecular dynamics (FPMD) codes that will enable the number of atoms simulated to be proportional to the number of processors available; with traditional algorithms, the size of simulations is much too small to model complex systems or realistic materials. The researchers have achieved excellent scaling on 100,000 cores of Vulcan with 100,000 atoms at a rate of about 4 minutes per time step.

Focus Area: 
Materials Science/Molecular Dynamics, Algorithm Development at Extreme Scale

By applying and extending ideas from data mining, image and video processing, statistics, and pattern recognition, we are developing a new generation of computational tools and techniques that are being used to improve the way in which scientists extract useful information from data.  Sapphire technology is being applied to problems in a variety of application areas including plasma physics experiments and simulations, remote sensing imagery, video surveillance, climate simulations, astronomy, and fluid mix experiments and simulations.

Focus Area: 
Data Analytics and Management, Scientific Data Management

SOAR (Stateless, One-pass Adaptive Refinement) is a view-dependent mesh refinement and rendering algorithm.

Focus Area: 
Visualization and Scientific Data Analysis

Sphinx, an integrated parallel microbenchmark suite, consists of a harness for running performance tests and extensive tests of MPI, Pthreads and OpenMP. LLNL adaptations include extensive tests of the Pthreads interface and integration of the LLNL OpenMP Performance Suite. Sphinx is highly portable, and the test harness provides a flexible mechanism for running performance tests. The entire suite is implemented in C and has been run on a wide variety of platforms.

Focus Area: 
Parallel Software Development Tools, MPI, OpenMP

Veritas provides a method for validating proxy applications to ensure that they capture the intended characteristics of their parents. Previously, the validation process has been done mostly by manually matching algorithmic steps in proxy applications to the parent or by relying on the experience of the code developer. Veritas can identify and compare performance sinks in areas such as memory, cache utilization, and network utilization.

Focus Area: 
Co-design

Researchers develop innovative data representations and algorithms to provide faster, more efficient ways to preserve information encoded in data.

Focus Area: 
Data Analytics and Management, Extreme Computing, Hardware Architecture, Memory-Centric Architectures, Mesh Management