Extending the Uintah Framework through the Petascale Modeling of Detonation in Arrays of High Explosive Devices

Abstract

The Uintah software framework for the solution of a broad class of fluid-structure interaction problems has been developed by using a problem-driven approach that dates back to its inception. Uintah uses a layered taskgraph approach that decouples the problem specification as a set of tasks from the adaptive runtime system that executes these tasks. Using this approach it is possible to improve the performance of the software components to enable the solution of broad classes of problems as well as the driving problem itself. This process is illustrated by a motivating problem that is the computational modeling of the hazards posed by thousands of explosive devices during a Deflagration to Detonation Transition (DDT) that occurred on Highway 6 in Utah. In order to solve this complex fluid-structure interaction problem at the required scale, substantial algorithmic and data structure improvements were needed to Uintah. These improvements enabled scalable runs for the target problem and provided the capability to model the transition to detonation. The solution to the target problem from these runs provided insight as to why the detonation happened, as well as demonstrating a possible remediation strategy that may have avoided detonation.

Packing Configurations of PBX‐9501 Cylinders to Reduce the Probability of a Deflagration to Detonation Transition (DDT)

“…gaseousp roducts at multiple initial temperatures andp ressures [17,18,19,20].T he DDT modelu tilizes am odified Ward, Son, and Brewster (WSB) burn model [18,20,21] to evaluate the mass conversion rate, the ViscoSCRAM constitutive model [22] to model the damage in the solid, and the JWL ++ simple reactive flow model [23] to describe detonation.T he commonly used JWL equation of state [ 23,24] was used for the solid explosivea nd the product gases. Detonation occurs in Uintah's DDT model whent he localized pressure is greater than the pressure threshold, 5.3 GPa [17,18].F urther details on the model can be found in references [ 17,18,19,20,25].T he Uintah framework has al ong history of high performance computing and has shown goods trong and weak scaling characteristics up to 512 Kc ores on DOE's Mira [26,27,28].U intah's strong scalability enabled us to run large 2D and full 3D simulations at high grid resolutions (2 mm). The reactionm odel has been validated at many resolutions including 2mm [ 19,29].…”

PsmArena: Partitioned shared memory for NUMA-awareness in multithreaded scientific applications

“…First, since OS-based approaches can only work with fixed-size pages and applications usually work with variable-size blocks, fragmentation can become significant. For example, structured-grid scientific applications such as Uintah [10] and JASMIN [11] usually partition the computational domain into small blocks called "patches" and memory is allocated per-patch. For a patch of size 20 20, an array to store 20 20 doubleprecision number will be allocated, which utilizes Second, to mitigate the fragmentation, heap managers will kick in and share the page among threads.…”

Scalable Data Management of the Uintah Simulation Framework for Next-Generation Engineering Problems with Radiation

“…This graph-based approach allows tasks to execute in a manner that efficiently overlaps communication and computation, and includes out-of-order execution of tasks (with respect to a topological sort) where possible. Using this task-based approach also allows for improved load balancing, as only nodes need to be considered, not individual cores [8].…”

“…This work demonstrates the efficacy of our approach by adapting the Uintah computational framework [8], a highly scalable asynchronous many-task (AMT) [7] runtime system, to use our I/O system and spatial transport sweeps within a large-eddy simulation (LES). This work is aimed at predicting the performance of a commercial 1000 MWe USC coal boiler, and has been considered as an ideal exascale candidate given that the spatial and temporal resolution requirements on physical grounds give rise to problems between 50 to 1000 times larger than those we can solve today.…”