LLNL uses some of the most powerful computers in the world, including the 125-petaflops supercomputing system, Sierra, to ensure the safety and reliability of the nation's aging nuclear deterrent. This page describes some of the computer codes that WCI scientists develop and use.
ALE3D is a 2D and 3D, arbitrary Lagrangian-Eulerian, massively-parallel, finite-element code that treats fluid and elastic-plastic response of materials on an unstructured grid of 3D hexagonal elements or 2D quadrilaterals. The major components of the code are explicit and implicit continuum-mechanics, thermal diffusion, and chemistry. An incompressible flow model also exists along with a 3D Magneto Hydro Dynamic (MHD) capability.
The picture to the left shows a simulated cube contains 93,000 crystals and is 300 µm across.
ARDRA is a 1D, 2D, and 3D deterministic (SN) neutron and gamma transport code. ARDRA demonstrated scalability on LLNL's IBM BG/Q system Sequoia using 1,572,864 MPI tasks with over 37.5 trillion unknowns, while achieving over 71% parallel efficiency.
The image to the left shows isosurfaces derived from material properties and displays a slab of the interior of a spherical criticality benchmark problem. The left plot displays the geometry description while the right displays the neutron flux on the same slab. Red and blue colors on the right image represent total neutron scalar flux values and the highest and lowest flux values, respectively, while the colors in the left image highlight different sectors of the geometry. Click on the image for a larger view.
AMTRAN (Adaptive Mesh TRANsport is a 1D, 2D and 3D deterministic (SN) neutron and gamma transport code that uses spatial and angular adaptive refinement. It is able to run efficiently on a wide range of platforms, from serial to massively parallel architectures.
The picture to the left shows an example of local refinement.
BLAST is an object-oriented high-order finite element ALE hydrodynamics research code that supports 2Dxy/2Drz/3D unstructured curvilinear meshes. BLAST is capable of modeling complex multi-material shock hydrodynamics on massively parallel computers using advanced high-order curvilinear finite element methods. The BLAST algorithm has high arithmetic intensity and is well suited for advanced heterogeneous computing architectures. [More information about BLAST]
The image to the left shows a multi-material Lagrangian hydrodynamics simulation using 8th order curvilinear finite elements in BLAST. Click on the image for a larger view.
Cheetah, an LLNL thermochemical computer code, is a convenient and accurate physics- and chemistry-based computational tool for predicting the detonation velocities and energy release properties of energetic materials such as explosives, propellants, and pyrotechnics. Cheetah employs advanced concepts and theories of fluids and solids at high pressures and temperatures to model the thermodynamics of explosion products that result from the detonation of modern condensed, energy-dense explosives. The results can be used for materials design, performance predictions and formulation optimization, as well as safety analyses, accident forensics, and even to study homemade and terrorist devices. WCI makes particular use of the Cheetah database, which currently includes a wide range of approximately 1,000 chemical ingredients and products. [More information about Cheetah]
CRETIN is a 1D, 2D, and 3D non-local thermodynamic equilibrium (NLTE) atomic kinetics/radiation transport code which follows the time evolution of atomic populations and photon distributions as radiation interacts with a plasma. It can provide detailed spectra for comparing with experimental diagnostics.
KIM3D is a parallel hybrid plasma simulation code that treats ions as kinetic particles and uses the Darwin limit of Maxwell's equations. It is used primarily for modeling the physics of magnetized collisionless shocks and high altitude weapons effects.
The picture to the left shows a collisionless shock disturbing the Earth's ambient magnetic field. Click on the image for a larger view.
Kull is a massively parallel simulation code developed to model high energy density physics applications. Kull can be steered using the Python programming language but its computational kernels are written predominantly in C++.
The picture to the left shows a temperature plot with spots on the hohlraum surface.
Mercury is a 1D, 2D and 3D massively parallel Monte Carlo neutron, gamma and charged particle transport code. It is capable of tracking particles on both user defined meshes as well as user defined geometric shapes, or a combination of the two.
Miranda is a radiation hydrodynamics code designed for large-eddy simulation of multicomponent flows with turbulent mixing. Additional physics packages include magneto-hydrodynamics, self-gravity, and thermonuclear fusion. The hydro package is based on tenth-order compact (Pade) schemes for spatial differencing, combined with fourth-order Runge-Kutta timestepping. Applications to date include interfacial instabilities (RT, RM, and KH), ICF implosions, ablation physics, supernovae, the LIFE chamber, and classical fluids experiments (e.g., drop tanks, shock tubes, etc.).
The image to the left shows a Rayleigh-Taylor instability calculated by Miranda. Click on the image for a larger view.
PMesh is a suite of mesh generation tools developed to set up problems for massively parallel Advanced Simulation and Computing (ASC) simulations. DRACO is an interactive code used for geometry creation, decomposition, mesh generation, and visualization. XENA is a stand-alone, massively parallel mesh generator that uses DRACO geometry as input.
The image to the left shows DRACO example meshes.
Spheral is a modeling tool aimed at the development of novel meshless numerical modeling techniques, particularly focused on hydrodynamics, strength, and damage modeling. Spheral currently incorporates fluid and solid modeling via SPH (Smoothed Particle Hydrodynamics) and ASPH (Adaptive Smoothed Particle Hydrodynamics). [More about Spheral]
The image to the left shows gas-gun-driven expansion and rupture of a steel tube. Click on image for a larger version.
Uncertainty quantification, or "UQ," is the quantitative characterization and reduction of uncertainty in computer applications through running very large suites of calculations to characterize the effects of minor differences in the systems. Sources of uncertainty are rife in the natural sciences and engineering fields. UQ uses statistical methods to determine likely outcomes. Sequoia was the first system to allow for the routine use of two-dimensional UQ studies at high resolution; it will also be capable of entry-level three-dimensional UQ studies. [More about UQ]
The image to the left shows an OSIRIS simulation on Sequoia, which reveals the interaction of a fast-ignition-scale laser with dense deuterium–tritium plasma.
VisIt is an interactive, parallel visualization and graphical analysis tool for viewing scientific data on Unix and PC platforms. VisIt can process data for many of the simulation codes developed by Advanced Simulation and Computing (ASC). For more information, stop by the VisIt Web page.
The image to the left shows a volume plot. VisIT allows the potential to visualize all the data within a three-dimensional volume by mapping scalar values to an opacity and color value.