Resistive memories enable dramatic energy reductions for neural algorithms. We propose a general purpose neural architecture that can accelerate many different algorithms and determine the device properties that will be needed to run backpropagation on the neural architecture. To maintain high accuracy, the read noise standard deviation should be less than 5% of the weight range. The write noise standard deviation should be less than 0.4% of the weight range and up to 300% of a characteristic update (for the datasets tested). Asymmetric nonlinearities in the change in conductance vs pulse cause weight decay and significantly reduce the accuracy, while moderate symmetric nonlinearities do not have an effect. In order to allow for parallel reads and writes the write current should be less than 100 nA as well.

Electrical power efficiency is a primary concern in designing modern HPC systems. Common strategies to improve CPU power efficiency rely on increased parallelism within a processor that is enabled both by an increase in the vector capabilities within the core and also the number of cores within a processor. Although many-core processors have been available for some time, achieving power-efficient performance has been challenging due to the offload model. Here, we evaluate performance of the molecular dynamics code LAMMPS on two new Intel®processors including the second generation many-core Intel®Xeon Phi™ processor that is available as a bootable CPU. We describe our approach to measure power consumption out-of-band and software optimizations necessary to achieve energy efficiency. We analyze benefits from Intel®Advanced Vector Extensions 512 instructions and demonstrate increased simulations rates with over 9X the CPU+DRAM power efficiency when compared to the unoptimized code on previous generation processors.

The RandomAccess benchmark as defined by the High Performance Computing Challenge (HPCC) tests the speed at which a machine can update the elements of a table spread across global system memory, as measured in billions (giga) of updates per second (GUPS). The parallel implementation provided by HPCC typically performs poorly on distributed-memory machines, due to updates requiring numerous small point-to-point messages between processors. We present an alternative algorithm which treats the collection of P processors as a hypercube, aggregating data so that larger messages are sent, and routing individual datums through dimensions of the hypercube to their destination processor. The algorithm's computation (the GUP count) scales linearly with P while its communication overhead scales as log2(P), thus enabling better performance on large numbers of processors. The new algorithm achieves a GUPS rate of 19.98 on 8192 processors of Sandia's Red Storm machine, compared to 1.02 for the HPCCprovided algorithm on 10350 processors. We also illustrate how GUPS performance varies with the benchmark's specification of its "look-ahead" parameter. As expected, parallel performance degrades for small look-ahead values, and improves dramatically for large values.

Discrete ordinates methods are commonly used to simulate radiation transport for fire or weapons modeling. The computation proceeds by sweeping the flux across a grid. A particular cell cannot be computed until all the cells immediately upwind of it are finished. If the directed dependence graph for the grid cells contains a cycle then sweeping methods will deadlock. This can happen in unstructured grids and time stepped problems where the grid is allowed to deform. In this paper we present a parallel algorithm to detect cycles in the dependence graphs present in these grids as well as an implementation and experimental results on shared and distributed memory machines.

The method of discrete ordinates is commonly used to solve the Boltzmann radiation transport equation for applications ranging from simulations of fires to weapons effects. The equations are most efficiently solved by sweeping the radiation flux across the computational grid. For unstructured grids this poses several interesting challenges, particularly when implemented on distributed-memory parallel machines where the grid geometry is spread across processors. We describe a asynchronous, parallel, message-passing algorithm that performs sweeps simultaneously from many directions across unstructured grids. We identify key factors that limit the algorithm’s parallel scalability and discuss two enhancements we have made to the basic algorithm: one to prioritize the work within a processor’s subdomain and the other to better decompose the unstructured grid across processors. Performance results are give for the basic and enhanced algorithms implemented withi a radiation solver running on hundreds of processors of Sandia’s Intel Tflops machine and DEC-Alpha CPlant cluster.

A number of computational procedures employ multiple grids on which solutions are computed. For example, in multi-physics simulations a primary grid may be used to compute mechanical deformation of an object while a secondary grid is used for thermal conduction calculations. When modeling coupled thermo-mechanical effects, solution data must be interpolated back and forth between the grids each timestep. On a parallel machine, this grid transfer operation can be challenging if the two grids are decomposed to processors differently for reasons of computational efficiency. If the grids move or adapt separately, the complexity of the operation is compounded. In this paper we describe a grid transfer algorithm suitable for massively parallel codes which use multiple grids. It uses a rendezvous technique wherein a third decomposition is used to search for elements in one grid that contain nodal points of the other. This has the advantage of enabling the grid transfer to be load-balanced separately from the remainder of the computations. The algorithm has been implemented as an object-oriented tool for the multi-physics code SIERRA, currently under development at Sandia. Performance and scalability results for the grid transfer operation running on the Intel/Sandia TFLOPS supercomputer are presented.

We briefly discuss three algorithms used to process synthetic aperture radar (SAR) images: a polar reformat­ ter that maps raw SAR data to a rectangular mesh, 2-d FFTs that the data set to a real image, and phase gradient autofocusing which creates clear images by cor­ recting blurs resulting from uncompensated phase errors. Implementation details and performance data for the algo­ rithms are given for two parallel supercomputers, a nCUBE 2 hypercube and a CM-2. Portions of the algo­ rithms execute at gigaflop speeds, meaning that large SAR data sets (16,384 x 65,536) can be processed in tens of minutes instead of hours as on conventional supercomput­ ers. Equally significantly, the gigabytes of internal mem­ ory on the parallel machines allow entire images to be