Search NASA⌕ Search

SEARCH · Search NASA

Results for “near memory computing”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 253 records · Page 14

Large-scale integration of artificial atoms in hybrid photonic circuits

A central challenge in developing quantum computers and long-range quantum networks is the distribution of entanglement across many individually controllable qubits. Colour centres in diamond have emerged as leading solid-state ‘artificial atom’ qubits because they enable on-demand remote entanglement, coherent control of over ten ancillae qubits with minute-long coherence times and memory-enhanced quantum communication. A critical next step is to integrate large numbers of artificial atoms with photonic architectures to enable large-scale quantum information processing systems. So far, these efforts have been stymied by qubit inhomogeneities, low device yield and complex device requirements. Here we introduce a process for the high-yield heterogeneous integration of ‘quantum microchiplets’—diamond waveguide arrays containing highly coherent colour centres—on a photonic integrated circuit (PIC). We use this process to realize a 128-channel, defect-free array of germanium-vacancy and silicon-vacancy colour centres in an aluminium nitride PIC. Photoluminescence spectroscopy reveals long-term, stable and narrow average optical linewidths of 54 megahertz (146 megahertz) for germanium-vacancy (silicon-vacancy) emitters, close to the lifetime-limited linewidth of 32 megahertz (93 megahertz). We show that inhomogeneities of individual colour centre optical transitions can be compensated in situ by integrated tuning over 50 gigahertz without linewidth degradation. The ability to assemble large numbers of nearly indistinguishable and tunable artificial atoms into phase-stable PICs marks a key step towards multiplexed quantum repeaters and general-purpose quantum processors

36 MATERIALS SCIENCE↗

Forecasting of Storm-Surge Floods Using ADCIRC and Optimized DEMs

Increasing the accuracy of storm-surge flood forecasts is essential for improving preparedness for hurricanes and other severe storms and, in particular, for optimizing evacuation scenarios. An interactive database, developed by WorldWinds, Inc., contains atlases of storm-surge flood levels for the Louisiana/Mississippi gulf coast region. These atlases were developed to improve forecasting of flooding along the coastline and estuaries and in adjacent inland areas. Storm-surge heights depend on a complex interaction of several factors, including: storm size, central minimum pressure, forward speed of motion, bottom topography near the point of landfall, astronomical tides, and, most importantly, maximum wind speed. The information in the atlases was generated in over 100 computational simulations, partly by use of a parallel-processing version of the ADvanced CIRCulation (ADCIRC) model. ADCIRC is a nonlinear computational model of hydrodynamics, developed by the U.S. Army Corps of Engineers and the US Navy, as a family of two- and three-dimensional finite-element-based codes. It affords a capability for simulating tidal circulation and storm-surge propagation over very large computational domains, while simultaneously providing high-resolution output in areas of complex shoreline and bathymetry. The ADCIRC finite-element grid for this project covered the Gulf of Mexico and contiguous basins, extending into the deep Atlantic Ocean with progressively higher resolution approaching the study area. The advantage of using ADCIRC over other storm-surge models, such as SLOSH, is that input conditions can include all or part of wind stress, tides, wave stress, and river discharge, which serve to make the model output more accurate. To keep the computational load manageable, this work was conducted using only the wind stress, calculated by using historical data from Hurricane Camille, as the input condition for the model. Hurricane storm-surge simulations were performed on an eight-node Linux computer cluster. Each node contained dual 2-GHz processors, 2GB of memory, and a 40GB hard drive. The digital elevation model (DEM) for this region was specified using a combination of Navy data (over water), NOAA data (for the coastline), and optimized Interferometric Synthetic Aperture Radar data (over land). This high-resolution topographical data of the Mississippi coastal region provided the ADCIRC model with improved input with which to calculate improved storm-surge forecasts.

Valenti, Elizabeth↗

ZFP Hardware Implementation

ZFP is a floating-point compression format gaining traction in high-performance computing applications. A software implementation has demonstrated the ability to reduce data movement across communication channels and to reduce the footprint of floating-point arrays in memory. Nevertheless, any benefit in performance is limited to the spare compute cycles available before reaching bandwidth limitations. A hardware implementation of ZFP has the potential to substantially improve HPC application performance. Hardware compression speedup over software alone on a single CPU core ranges from 15x for 1-D arrays to over 200x for 3-D arrays. The implementation of ZFP is sourced in SystemC to facilitate its evaluation in various architectures. The encode pipeline consists of several modules. Uncompressed blocks of floating-point numbers in IEEE format flow into the pipeline in a stream and a compressed bitstream flows out from the pipeline. The decode pipeline is the inverse of the encode pipeline. A modular design enables formats other than IEEE (such as posits) to be considered with minor adaptations. Best performance will be realized from the hardware ZFP unit when batches of blocks are processed at a time. The implementation supports 1-D, 2- D and 3-D blocks of floating-point numbers. The hardware implementation of ZFP has been validated with the software implementation of ZFP. C++ template parameters are used to specify the bit width of floating-point numbers and the array dimension of the encoder. A test bench program has been created with several test cases, some with continuous data that is ideal for ZFP, and others with extreme cases containing numbers near or at the maximum or minimum values supported by the number format.

Lloyd, GS↗

Approximate Weighted Matching On Emerging Manycore and Multithreaded Architectures

Graph matching is a prototypical combinatorial problem with many applications in computer science and scientific computing, but algorithms for computing optimal matchings are challenging to parallelize. Approximate matching algorithms provide an alternate route for parallelization, and in many contexts compute near-optimal matchings for large-scale graphs. We present sharedmemory parallel implementations for computing half-approximate weighted matching on state-of-the-art multicore (Intel Nehalem and AMD Magny-Cours), manycore (Nvidia Tesla and Nvidia Fermi) and massively multithreaded (Cray XMT) platforms. We provide two implementations: the first implementation uses shared work queues, and is suited to all these platforms; the second implementation is based on dataflow principles, and exploits the architectural features of the Cray XMT. Using a carefully chosen dataset that exhibits characteristics from a wide range of real-world applications, we show scalable performance across different platforms. In particular, for one instance of the input, an R-MAT graph (RMAT-G), we show speedups of: about 32 on 48 cores of an AMD Magny-Cours; 7 on 8 cores of Intel Nehalem; 3 on Nvidia Tesla and 10 on Nvidia Fermi relative to one core of Intel Nehalem; and 60 on 128 processors of Cray XMT. We demonstrate good weak and strong scaling for graphs with up to a billion edges using up to 12, 800 threads. Given the breadth of this work, we focus on simplicity and portability of software rather than excessive fine-tuning for each platform. To the best of our knowledge, this is the first such large-scale study of the half-approximate weighted matching problem on shared-memory platforms. Driven by the critical enabling role of combinatorial algorithms such as matching in scientific computing and the emergence of informatics applications, there is a growing demand to support irregular computations on current and future computing platforms. In this context, we evaluate the capability of emerging multithreaded platforms to tolerate latency induced by irregular memory access patterns, and to support fine-grained parallelism via light-weight synchronization mechanisms. By contrasting the architectural features of these platforms against the Cray XMT, which is specifically designed to support irregular memory-intensive applications, we delineate the impact of these choices on performance.

Halappanavar, Mahantesh↗

Performance of the OVERFLOW-MLP and LAURA-MLP CFD Codes on the NASA Ames 512 CPU Origin System

The shared memory Multi-Level Parallelism (MLP) technique, developed last year at NASA Ames has been very successful in dramatically improving the performance of important NASA CFD codes. This new and very simple parallel programming technique was first inserted into the OVERFLOW production CFD code in FY 1998. The OVERFLOW-MLP code's parallel performance scaled linearly to 256 CPUs on the NASA Ames 256 CPU Origin 2000 system (steger). Overall performance exceeded 20.1 GFLOP/s, or about 4.5x the performance of a dedicated 16 CPU C90 system. All of this was achieved without any major modification to the original vector based code. The OVERFLOW-MLP code is now in production on the inhouse Origin systems as well as being used offsite at commercial aerospace companies. Partially as a result of this work, NASA Ames has purchased a new 512 CPU Origin 2000 system to further test the limits of parallel performance for NASA codes of interest. This paper presents the performance obtained from the latest optimization efforts on this machine for the LAURA-MLP and OVERFLOW-MLP codes. The Langley Aerothermodynamics Upwind Relaxation Algorithm (LAURA) code is a key simulation tool in the development of the next generation shuttle, interplanetary reentry vehicles, and nearly all "X" plane development. This code sustains about 4-5 GFLOP/s on a dedicated 16 CPU C90. At this rate, expected workloads would require over 100 C90 CPU years of computing over the next few calendar years. It is not feasible to expect that this would be affordable or available to the user community. Dramatic performance gains on cheaper systems are needed. This code is expected to be perhaps the largest consumer of NASA Ames compute cycles per run in the coming year.The OVERFLOW CFD code is extensively used in the government and commercial aerospace communities to evaluate new aircraft designs. It is one of the largest consumers of NASA supercomputing cycles and large simulations of highly resolved full aircraft are routinely undertaken. Typical large problems might require 100s of Cray C90 CPU hours to complete. The dramatic performance gains with the 256 CPU steger system are exciting. Obtaining results in hours instead of months is revolutionizing the way in which aircraft manufacturers are looking at future aircraft simulation work. Figure 2 below is a current state of the art plot of OVERFLOW-MLP performance on the 512 CPU Lomax system. As can be seen, the chart indicates that OVERFLOW-MLP continues to scale linearly with CPU count up to 512 CPUs on a large 35 million point full aircraft RANS simulation. At this point performance is such that a fully converged simulation of 2500 time steps is completed in less than 2 hours of elapsed time. Further work over the next few weeks will improve the performance of this code even further.The LAURA code has been converted to the MLP format as well. This code is currently being optimized for the 512 CPU system. Performance statistics indicate that the goal of 100 GFLOP/s will be achieved by year's end. This amounts to 20x the 16 CPU C90 result and strongly demonstrates the viability of the new parallel systems rapidly solving very large simulations in a production environment.

Taft, James R.↗

Analysis of vibronic coupling in a 4f molecular magnet with FIRMS

Vibronic coupling, the interaction between molecular vibrations and electronic states, is a fundamental effect that profoundly affects chemical processes. In the case of molecular magnetic materials, vibronic, or spin-phonon, coupling leads to magnetic relaxation, which equates to loss of magnetic memory and loss of phase coherence in molecular magnets and qubits, respectively. The study of vibronic coupling is challenging, and most experimental evidence is indirect. Here we employ far-infrared magnetospectroscopy to directly probe vibronic transitions in [Yb(trensal)] (where H 3 trensal = 2,2,2-tris(salicylideneimino)trimethylamine). We find intense signals near electronic states, which we show arise due to an “envelope effect” in the vibronic coupling Hamiltonian, which we calculate fully ab initio to simulate the spectra. We subsequently show that vibronic coupling is strongest for vibrational modes that simultaneously distort the first coordination sphere and break the C 3 symmetry of the molecule. With this knowledge, vibrational modes could be identified and engineered to shift their energy towards or away from particular electronic states to alter their impact. Hence, these findings provide new insights towards developing general guidelines for the control of vibronic coupling in molecules.

36 MATERIALS SCIENCE↗

Helium: lifting high-performance stencil kernels from stripped x86 binaries to halide DSL code

Highly optimized programs are prone to bit rot, where performance quickly becomes suboptimal in the face of new hardware and compiler techniques. In this paper we show how to automatically lift performance-critical stencil kernels from a stripped x86 binary and generate the corresponding code in the high-level domain-specific language Halide. Using Halide's state-of-the-art optimizations targeting current hardware, we show that new optimized versions of these kernels can replace the originals to rejuvenate the application for newer hardware. The original optimized code for kernels in stripped binaries is nearly impossible to analyze statically. Instead, we rely on dynamic traces to regenerate the kernels. We perform buffer structure reconstruction to identify input, intermediate and output buffer shapes. Here, we abstract from a forest of concrete dependency trees which contain absolute memory addresses to symbolic trees suitable for high-level code generation. This is done by canonicalizing trees, clustering them based on structure, inferring higher-dimensional buffer accesses and finally by solving a set of linear equations based on buffer accesses to lift them up to simple, high-level expressions. Helium can handle highly optimized, complex stencil kernels with input-dependent conditionals. We lift seven kernels from Adobe Photoshop giving a 75 % performance improvement, four kernels from Irfan View, leading to 4.97 x performance, and one stencil from the mini GMG multigrid benchmark netting a 4.25 x improvement in performance. We manually rejuvenated Photoshop by replacing eleven of Photoshop's filters with our lifted implementations, giving 1.12 x speedup without affecting the user experience.

97 MATHEMATICS AND COMPUTING↗

The organic redox transistor for neuromorphic computing

Inspired by the in-memory computing architectures of biological systems, neuromorphic computing using crossbar arrays of artificial synapses based on non-volatile memory (NVM) devices with variable conductances has emerged as a new paradigm to enable massively parallel and ultra-low power computing hardware for data centric applications. Although inference has been demonstrated successfully using crossbars based on a variety of NMV technologies, efficient learning and scaling to large arrays (>10 6 elements) remains a challenge due to the synaptic elements' non-ideal electrical characteristics which degrades ANN accuracy. A further challenge is that in the conductive state memristors draw large currents >μA resulting in significant voltage drops in the interconnect wires and increased probability of failure in scaled arrays. We suggest the organic polymer redox transistor (RT) is an alternate approach that could solve many of these challenges, enabling both inference and parallel outer product updates, as recently demonstrated by Fuller et al. An RT consists of redox-active channel and gate electrodes in contact with a liquid or solid electrolyte. lon insertion through the electrolyte controls the channel electronic conductivity, while electron transfer through an external circuit maintains overall charge neutrality. Unlike a rechargeable battery, in the RT the voltage built-up across the electrolyte is kept to a minimum (typically <100 mV) by using the same material for the gate and channel. Elimination of the voltage offset simplifies integration of the RT into programmable arrays by enabling the use of various selectors. RTs based on inorganic and organic materials have been recently demonstrated with conductance tuning occurring at potentials of just a few mV and hundreds to thousands of linearly and symmetrically programmable conductance states, enabling near ideal accuracy in neural network simulations. Introduced in the 1980's, redox transistors with metallic gate electrodes and organic channel materials, also known as organic electrochemical transistors (OECTs), have been explored for a variety of applications such as chem- and bio-sensing, neural interfaces, and low cost printed circuits. A typical channel material for OECTs is the conducting polymer poly(3,4-ethylenedioxythiophene) doped with poly(styrene sulfonate) (PEDOT:PSS). PEDOT is a p-type semiconducting polymer with mobile positively charged polarons that hop chain-to-chain.

Talin, Albert Alec↗

Electronic Structure of Buried Interfaces - Oral Presentation [Slides]

In the electronics behind computer memory storage, the speed and size are dictated by the performance of permanent magnets inside devices called read heads. Complicated magnets made of stacked layers of thin films can be engineered to have properties that yield more energy storage and faster switching times compared to conventional iron or cobalt magnets. The reason is that magnetism is a result of subtle interactions amongst electrons; just how neurons come together on large scales to make cat brains and dog brains, ensembles of electrons interact and become ferromagnets and paramagnets. These interactions make magnets too difficult to study in their entirety, so I focus on the interfaces between layers, which are responsible for the coupling materials physicists hope to exploit to produce next-generation magnets. This project, I study a transition metal oxide material called LSCO, Lanthanum Cobaltite, which can be a paramagnet or a ferromagnet depending on how you tweak the electronic structure. It exhibits an exciting behavior: its sum is greater than the sum of its parts. When another similar material called a LSMO, Lanthanum Manganite, is grown on top of it, their interface has a different type of magnetism from the LSCO or the LSMO! I hope to explain this by demonstrating differently charged ions in the interface. The typical method for quantifying this is x-ray absorption, but all conventional techniques look at every layer simultaneously, averaging the interfaces and the LSCO layers that we want to characterize separately. Instead, I must use a new reflectivity technique, which tracks the intensity of reflected x-rays at different angles, at energies near the absorption peaks of certain elements, to track changes in the electronic structure of the material. The samples were grown by collaborators at the Takamura group at U.C. Davis and probed with this “resonant reflectivity” technique on Beamline 2-1 at the Stanford Synchrotron Radiation Lightsource. This project was funded by the Department of Energy and supported by the SLAC National Accelerator Laboratory. Preliminary results suggest that different ionic charges are indeed responsible for the different magnetic properties at the interface, which is promising because charge is easy to control. Once scientists understand how to tune the magnetic properties of materials like LSCO and LSMO, industries get closer to designing the magnets that will soon revolutionize consumer electronics.

36 MATERIALS SCIENCE↗

Speed and Memory Improvements to MCNP6 Delayed-Gamma Line Treatment

The delayed-gamma (DG) line-emission (line) sampling was first introduced into MCNPX in 2005. At the time, storage demands for problems with large numbers of DG lines was a major concern. In order to accommodate large amounts of line data while keeping the memory footprint of such a problem from exceeding memory limits, complicated logic was included to sift through various storage options. In mid-2015 concerns arose over the adequacy of the MCNP6 DG source option (inherited from MCNPX) with line sampling for modeling high-fidelity problems. One indication that there was a problem came when studying the MCNP6 produced DG spectra of U-238 and its progeny using high-fidelity spectroscopic tallies when compared to DG line amplitudes directly from ENDF/B-VII.1 emission data. Despite producing a correct integral number of gammas, the comparison to ENDF data showed a broadening in the energy lines and a reduction in the amplitude of more prominent lines. It was later found that the cause of this broadening was one of the memory-saving techniques used to store DG line data when problems required many DG lines. Specifically, the DG lines were integrated into fine energy bins for sampling. When a DG line was to be sampled, a random number would be used to choose which energy bin to sample based on emission probability. Once chosen, a second random number was used to perform nearest-line sampling in that bin. The integration of lines led to an oversampling of low-probability lines, especially when in close proximity to high-probability lines. This treatment also caused a similar issue with the time-dependent behavior of DG line emission which occurred when low-probability lines were sampled near high probability lines with the former decay constant being used. The effect on time-dependent gamma production was demonstrated in analytic benchmarks performed by Weldon et al.. To address these issues a new gamma line storage and sampling treatment was introduced in 2015. This new treatment, stores gamma emission data in a line-by-line CDF which is sampled directly, thereby resolving both the line-broadening issue, and the time-dependence issue as was demonstrated by Tutt el al.. In addition, the new DG sampling treatment was written to use dynamically dimensioned storage arrays as opposed to large fixed-dimensioned arrays, as in the former treatment, and also makes more efficient use of line data that is stored by reducing or eliminating repeated data. As a result, storage arrays are generated only for DG data relevant to the problem being executed. The result should be a reduction in both memory requirements and runtime. However, these reductions could not be realized until the logic for the old treatment could be removed. This paper provides an update on recent work which has focused on the removal of the old DG sampling logic and attempts to quantify the memory reduction and speed up of selected problems.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NASA Tech Briefs, December 2009

Topics include: A Deep Space Network Portable Radio Science Receiver; Detecting Phase Boundaries in Hard-Sphere Suspensions; Low-Complexity Lossless and Near-Lossless Data Compression Technique for Multispectral Imagery; Very-Long-Distance Remote Hearing and Vibrometry; Using GPS to Detect Imminent Tsunamis; Stream Flow Prediction by Remote Sensing and Genetic Programming; Pilotless Frame Synchronization Using LDPC Code Constraints; Radiometer on a Chip; Measuring Luminescence Lifetime With Help of a DSP; Modulation Based on Probability Density Functions; Ku Telemetry Modulator for Suborbital Vehicles; Photonic Links for High-Performance Arraying of Antennas; Reconfigurable, Bi-Directional Flexfet Level Shifter for Low-Power, Rad-Hard Integration; Hardware-Efficient Monitoring of I/O Signals; Video System for Viewing From a Remote or Windowless Cockpit; Spacesuit Data Display and Management System; IEEE 1394 Hub With Fault Containment; Compact, Miniature MMIC Receiver Modules for an MMIC Array Spectrograph; Waveguide Transition for Submillimeter-Wave MMICs; Magnetic-Field-Tunable Superconducting Rectifier; Bonded Invar Clip Removal Using Foil Heaters; Fabricating Radial Groove Gratings Using Projection Photolithography; Gratings Fabricated on Flat Surfaces and Reproduced on Non-Flat Substrates; Method for Measuring the Volume-Scattering Function of Water; Method of Heating a Foam-Based Catalyst Bed; Small Deflection Energy Analyzer for Energy and Angular Distributions; Polymeric Bladder for Storing Liquid Oxygen; Pyrotechnic Simulator/Stray-Voltage Detector; Inventions Utilizing Microfluidics and Colloidal Particles; RuO2 Thermometer for Ultra-Low Temperatures; Ultra-Compact, High-Resolution LADAR System for 3D Imaging; Dual-Channel Multi-Purpose Telescope; Objective Lens Optimized for Wavefront Delivery, Pupil Imaging, and Pupil Ghosting; CMOS Camera Array With Onboard Memory; Quickly Approximating the Distance Between Two Objects; Processing Images of Craters for Spacecraft Navigation; Adaptive Morphological Feature-Based Object Classifier for a Color Imaging System; Rover Slip Validation and Prediction Algorithm; Safety and Quality Training Simulator; Supply-Chain Optimization Template; Algorithm for Computing Particle/Surface Interactions; Cryogenic Pupil Alignment Test Architecture for Aberrated Pupil Images; and Thermal Transport Model for Heat Sink Design.

Source record↗

Serpentine: Finite Difference Methods for Wave Propagation in Second Order Formulation

Wave propagation phenomena are important in many DOE applications such as nuclear explosion monitoring, geophysical exploration, estimating ground motion hazards and damage due to earthquakes, non-destructive testing, underground facilities detection, and acoustic noise propagation. There are also future applications that would benefit from simulating wave propagation, such as geothermal energy applications and monitoring sites for carbon storage via seismic reflection techniques. In acoustics and seismology, it is of great interest to increase the frequency bandwidth in simulations. In seismic exploration, greater frequency resolution enables shorter wave lengths to be included in the simulations, allowing for better resolution in the seismic imaging. In nuclear explosion monitoring, higher frequency seismic waves are essential for accurate discrimination between explosions and earthquakes. When simulating earthquake induced motion of large structures, such as nuclear power plants or dams, increased frequency resolution is essential for realistic damage predictions. Another example is simulations of micro-seismic activity near geothermal energy plants. Here, hydro-fracturing induces many small earthquakes and the time scale of each event is proportional to the square root of the moment magnitude. As a result, the motion is dominated by higher frequencies for smaller seismic events. The above wave propagation problems are all governed by systems of hyperbolic partial differential equations in second order differential form, i.e., they contain second order partial derivatives of the dependent variables. Our general research theme in this project has been to develop numerical methods that directly discretize the wave equations in second order differential form. The obvious advantage of working with hyperbolic systems in second order differential form, as opposed to rewriting them as first order hyperbolic systems, is that the number of differential equations in the second order system is significantly smaller. Another issue with re-writing a second order system into first order form is that compatibility conditions often must be imposed on the first order form. These (Saint-Venant) conditions ensure that the solution of the first order system also satisfies the original second order system. However, such conditions can be difficult to enforce on the discretized equations, without introducing additional modeling errors. This project has previously developed robust and memory efficient algorithms for wave propagation including effects of curved boundaries, heterogeneous isotropic, and viscoelastic materials. Partially supported by internal funding from Lawrence Livermore National Laboratory, many of these methods have been implemented in the open source software WPP, which is geared towards 3-D seismic wave propagation applications. This code has shown excellent scaling on up to 32,768 processors and has enabled seismic wave calculations with up to 26 Billion grid points. The WPP calculations have resulted in several publications in the field of computational seismology, e.g.. All of our current methods are second order accurate in both space and time. The benefits of higher order accurate schemes for wave propagation have been known for a long time, but have mostly been developed for first order hyperbolic systems. For second order hyperbolic systems, it has not been known how to make finite difference schemes stable with free surface boundary conditions, heterogeneous material properties, and curvilinear coordinates. The importance of higher order accurate methods is not necessarily to make the numerical solution more accurate, but to reduce the computational cost for obtaining a solution within an acceptable error tolerance. This is because the accuracy in the solution can always be improved by reducing the grid size h. However, in practice, the available computational resources might not be large enough to solve the problem with a low order method.

15 GEOTHERMAL ENERGY↗

Quantitative Performance Assessment of Proxy Apps and Parents (ECP Proxy App Project Milestone ADCD-504-9)

This report presents highlights of these efforts. Section 2 describes work that has been done to compare the performance of proxy applications on AMD MI60 vs. Nvidia V100 GPUs. So far only a small set of ECP proxies are running on AMD GPUs, but we will continue to expand this analysis as additional proxies become available. We find that although the MI60 and V100 have nearly the same measured memory bandwidth, memory bound proxy app kernels perform 20-30% worse on the MI60. Further work is needed to refine these comparisons to determine whether the root cause is due to differences in the hardware, software stack, platform specific optimization, or some combination of the three. Section 3 describes our continuing effort to find methods to accurately assess the similarity of proxies and parents. We have recently seen very encouraging results using a cosine similarity metric. This technique uses the angle between two vectors of hardware performance counters to characterize the similarity (or difference) between two applications or proxies. We show not only that several widely used proxies are highly similar to their parents, but also that they differ from non-related codes. We also show that cosine similarity can be used to identify gaps and redundancies in suites and even to gain insight into the effects of architectural differences between platforms. Our work on assessing the Exascale toolchain is ongoing. Our successes with performance measurement tools are evident from the data provided in this report. However, our assessments across the broader tool chain are still too incomplete to provide a meaningful report at this time. We will continue to assess tools and work with vendors and third party developers as issues are identified.

97 MATHEMATICS AND COMPUTING↗

A performance study of the time-varying cache behavior: a study on APEX, Mantevo, NAS, and PARSEC

Cache has long been used to minimize the latency of main memory accesses by storing frequently used data near the processor. Processor performance depends on the underlying cache performance. Therefore, significant research has been done to identify the most crucial metrics of cache performance. Although the majority of research focuses on measuring cache hit rates and data movement as the primary cache performance metrics, cache utilization is significantly important. We investigate the application’s locality using cache utilization metrics. In addition, we present cache utilization and traditional cache performance metrics as the program progresses providing detailed insights into the dynamic application behavior on parallel applications from four benchmark suites running on multiple cores. We explore cache utilization for APEX, Mantevo, NAS, and PARSEC, mostly scientific benchmark suites. Our results indicate that 40% of the data bytes in a cache line are accessed at least once before line eviction. Also, on average a byte is accessed two times before the cache line is evicted for these applications. Moreover, we present runtime cache utilization, as well as, conventional performance metrics that illustrate a holistic understanding of cache behavior. To facilitate this research, we build a memory simulator incorporated into the Structural Simulation Toolkit (Rodrigues et al. in SIGMETRICS Perform Eval Rev 38(4):37–42, 2011). Finally, our results suggest that variable cache line size can result in better performance and can also conserve power.

97 MATHEMATICS AND COMPUTING↗

DNS and LES of turbulent flow in a closed channel featuring a pattern of hemispherical roughness elements

Three mesh adaptivity algorithms were developed to facilitate and expedite the use of the CADIS and FW-CADIS hybrid Monte Carlo/deterministic techniques in accurate full-scale neutronics simulations of fusion energy systems with immense sizes and complicated geometries. First, a macromaterial approach enhances the fidelity of the deterministic models without changing the mesh. Second, a deterministic mesh refinement algorithm generates meshes that capture as much geometric detail as possible without exceeding a specified maximum number of mesh elements. Finally, a weight window coarsening algorithm decouples the weight window mesh and energy bins from the mesh and energy group structure of the deterministic calculations in order to remove the memory constraint of the weight window map from the deterministic mesh resolution. The three algorithms were used to enhance an FW-CADIS calculation of the prompt dose rate throughout the ITER experimental facility and resulted in a 23.3% increase in the number of mesh tally elements in which the dose rates were calculated in a 10-day Monte Carlo calculation. Additionally, because of the significant increase in the efficiency of FW-CADIS simulations, the three algorithms enabled this difficult calculation to be accurately solved on a regular computer cluster, eliminating the need for a world-class super computer.

Nuclear Science & Technology↗

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count. Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 2048 3 on a single node to 32768 3 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 2048 3 where communication is entirely intra-node, and for 32768 3 , where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 32768 3 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

3D fast Fourier transform↗

Development of Slat-gap and Slat-cove Filler Treatments for Noise Reduction Assessment on the High Lift Common Research Model in the NASA LaRC 14x22

Airframe noise is an important issue during the approach and landing phases of flight because engines are throttled back, and airframe sources become an equal or greater contributor to environmental noise. The leading-edge slat component of the high-lift system for transport aircraft has been identified as a prominent source of airframe noise. Noise production mechanisms and potential treatments to mitigate airframe noise have been areas of active research for nearly three decades. Two promising technologies for slat noise reduction are the slat-gap filler (SGF) and slat-cove filler (SCF). Previous work was done using a 2D section of a transport class aircraft wing to demonstrate their feasibility via benchtop and computational models. A wind tunnel test campaign was recently undertaken to test the SGF and SCF technologies in a flight-like environment by installing dynamically scaled treatments on the 10%-scale High-Lift Common Research Model (CRM-HL), which is representative of a commercial twin-aisle transport aircraft. Advancement of the SGF and SCF technologies to overcome challenges of integration with 3D airframe structures, computational simulations of the CRM-HL predicting the aerodynamic and aeroacoustic performance and structural loads, and design of wind tunnel model hardware with features accommodating noise treatments were conducted in parallel. Details of innovations in the SGF and SCF technologies for 3D airframe feasibility will be described. The loads the treatments must sustain in flight and the influence of those loads on the design is also discussed. Finally, the design of wing leading-edge assemblies to treat the 10% CRM-HL with dynamically scaled SGF and SCF components is presented. The treatments were successfully tested in 2020/2021 providing a demonstration of leading-edge noise reduction using components that would allow slat articulation.

airframe noise reduction↗

Scalla: Structured Cluster Architecture for Low Latency Access

Scalla is a distributed low-latency file access system that incorporates novel techniques that minimize latency and maximize scalability over a large distributed system with a distributed namespace. Scalla's techniques have shown to be effective in nearly a decade of service for the high-energy physics community using commodity hardware and interconnects. We describe the two components used in Scalla that are instrumental in its ability to provide low-latency, fault-tolerant name resolution and load distribution, and enable its use as a high-throughput, low-latency communication layer in the Qserv system, the Large Synoptic Survey Telescope's (LSST's) prototype astronomical query system. Scalla arguably exceeded its three main design objectives: low latency, scaling, and recoverability. In retrospect, these objectives were met using a simple but effective design. Low latency was met by uniformly using linear or constant time algorithms in all high-use paths, avoiding locks whenever possible, and using compact data structures to maximize the memory caching efficiency. Scaling was achieved by architecting the system as a 64-ary tree. Nodes can be added easily and as the number of nodes increases, search performance increases at an exponential rate. Recoverability is inherent in that no permanent state information is maintained and whatever state information is needed it can be quickly constructed or reconstructed in real time. This allows dynamic changes in a cluster of servers with little impact on over-all performance or usability. Today, Scalla is being deployed in environments and for uses that were never conceived in 2001. This speaks well for the systems adaptability but the underlying reason is that the system can meet its three fundamental objectives at the same time.

99 GENERAL AND MISCELLANEOUS↗