Kokkos: Enabling manycore performance portability through polymorphic memory access patterns

3-D Parachute simulation by the immersed boundary method

The manycore revolution in computational hardware can be characterized by increasing thread counts, decreasing memory per thread, and architecture specific performance constraints for memory access patterns. High performance computing (HPC) on emerging many core architectures requires codes to exploit every opportunity for thread-level parallelism and satisfy conflicting performance constraints. We developed the Kokkos C++ library to provide scientific and engineering codes with a user accessible many core performance portable programming model. The two foundational abstractions of Kokkos are (1) dispatch work to a many core device for parallel execution and (2) manage multidimensional arrays with polymorphic layouts. The integration of these abstractions enables users' code to satisfy multiple architecture specific memory access pattern performance constraints without having to modify their source code. In this paper we describe the Kokkos abstractions, summarize its application programmer interface (API), and present performance results for a molecular dynamics computational kernel and finite element mini-application.

/pdf/kokkos-enabling-performance-portability-across-manycore-25e0ud21m7.pdf

Kokkos: Enabling Performance Portability Across Manycore Architectures

In this paper, aerodynamic decelerators are defined as textile devices intended to be deployed at Mach numbers below five. Such aerodynamic decelerators include parachutes and inflatable aerodynamic decelerators (often known as ballutes). Aerodynamic decelerators play a key role in the Entry, Descent, and Landing (EDL) of planetary exploration vehicles. Among the functions performed by aerodynamic decelerators for such vehicles are deceleration (often from supersonic to subsonic speeds), minimization of descent rate, providing specific descent rates (so that scientific measurements can be obtained), providing stability (drogue function - either to prevent aeroshell tumbling or to meet instrumentation requirements), effecting further aerodynamic decelerator system deployment (pilot function), providing differences in ballistic coefficients of components to enable separation events, and providing height and timeline to allow for completion of the EDL sequence. Challenging aspects in the development of aerodynamic decelerators for planetary exploration missions include: deployment in the unusual combination of high Mach numbers and low dynamic pressures, deployment in the wake behind a blunt-body entry vehicle, stringent mass and volume constraints, and the requirement for high drag and stability. Furthermore, these aerodynamic decelerators must be qualified for flight without access to the exotic operating environment where they are expected to operate. This paper is an introduction to the development and application of aerodynamic decelerators for robotic planetary exploration missions (including Earth sample return missions) from the earliest work in the 1960s to new ideas and technologies with possible application to future missions. An extensive list of references is provided for additional study.

/pdf/aerodynamic-decelerators-for-planetary-exploration-past-1l33fyroao.pdf

Aerodynamic Decelerators for Planetary Exploration: Past, Present, and Future

Recent work in metal additive manufacturing (AM) suggests that mechanical properties may vary with feature size; however, these studies do not provide a statistically robust description of this phenomenon, nor do they provide a clear causal mechanism. Because of the huge design freedom afforded by 3D printing, AM parts typically contain a range of feature sizes, with particular interest in smaller features, so the size effect must be well understood in order to make informed design decisions. This work investigates the effect of feature size on the stochastic mechanical performance of laser powder bed fusion tensile specimens. A high-throughput tensile testing method was used to characterize the effect of specimen size on strength, elastic modulus and elongation in a statistically meaningful way. The effective yield strength, ultimate tensile strength and modulus decreased strongly with decreasing specimen size: all three properties were reduced by nearly a factor of two as feature dimensions were scaled down from 6.25 mm to 0.4 mm. Hardness and microstructural observations indicate that this size dependence was not due to an intrinsic change in material properties, but instead the effects of surface roughness on the geometry of the specimens. Finite element analysis using explicit representations of surface topography shows the critical role surface features play in creating stress concentrations that trigger deformation and subsequent fracture. The experimental and finite element results provide the tools needed to make corrections in the design process to more accurately predict the performance of AM components.

Size-dependent stochastic tensile properties in additively manufactured 316L stainless steel

Large, complex scientific and engineering application code have a significant investment in computational kernels to implement their mathematical models. Porting these computational kernels to the collection of modern manycore accelerator devices is a major challenge in that these devices have diverse programming models, application programming interfaces APIs, and performance requirements. The Kokkos Array programming model provides library-based approach to implement computational kernels that are performance-portable to CPU-multicore and GPGPU accelerator devices. This programming model is based upon three fundamental concepts: 1 manycore compute devices each with its own memory space, 2 data parallel kernels and 3 multidimensional arrays. Kernel execution performance is, especially for NVIDIA® devices, extremely dependent on data access patterns. Optimal data access pattern can be different for different manycore devices --potentially leading to different implementations of computational kernels specialized for different devices. The Kokkos Array programming model supports performance-portable kernels by 1 separating data access patterns from computational kernels through a multidimensional array API and 2 introduce device-specific data access mappings when a kernel is compiled. An implementation of Kokkos Array is available through Trilinos [Trilinos website, http://trilinos.sandia.gov/, August 2011].

/pdf/manycore-performance-portability-kokkos-multidimensional-4urvxt3gp3.pdf

Manycore performance-portability: Kokkos multidimensional array library

In this paper the authors present the current status of an unsteady 3D parachute simulation code which is being developed at Sandia National Laboratories under the Department of Energy's Accelerated Strategic Computing Initiative (ASCI). The Vortex Inflation PARachute code (VIPAR) which embodies this effort will eventually be able to perform complete numerical simulations of ribbon parachute deployment, inflation, and steady descent. At the present time they have a working serial version of the uncoupled fluids code which can simulate unsteady 3D incompressible flows around bluff bodies made up of triangular membrane elements. A parallel version of the code has just been completed which will allow one to compute flows over complex geometries utilizing several thousand processors on one of the new DOE teraFLOP computers.

/pdf/on-the-development-of-a-gridless-inflation-code-for-1hh59xr4d6.pdf

On the Development of a Gridless Inflation Code for Parachute Simulations

Fluid-structure coupling for lightweight flexible bodies.

/pdf/on-the-development-of-a-gridless-inflation-code-for-2op9ax0i0x.pdf

On the development of a gridless inflation code for parachute simulations

This report describes a 3-D fluid mechanics code for predicting flow past bluff bodies whose surfaces can be assumed to be made up of shell elements that are simply connected. Version 1.0 of the VIPAR code (Vortex Inflation PARachute code) is described herein. This version contains several first order algorithms that we are in the process of replacing with higher order ones. These enhancements will appear in the next version of VIPAR. The present code contains a motion generator that can be used to produce a large class of rigid body motions. The present code has also been fully coupled to a structural dynamics code in which the geometry undergoes large time dependent deformations. Initial surface geometry is generated from triangular shell elements using a code such as Patran and is written into an ExodusII database file for subsequent input into VIPAR. Surface and wake variable information is output into two ExodusII files that can be post processed and viewed using software such as EnSight{trademark}.

/pdf/a-3-d-vortex-code-for-parachute-flow-predictions-vipar-d2w2ueisiv.pdf

Vicki Porter

Papers

Manycore performance-portability: Kokkos multidimensional array library

On the Development of a Gridless Inflation Code for Parachute Simulations

Fluid-structure coupling for lightweight flexible bodies.

On the development of a gridless inflation code for parachute simulations

A 3-D Vortex Code for Parachute Flow Predictions: VIPAR Version 1.0