SEARCH · Search NASA
Results for “memory efficiency”
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.
Memory Seeds Enable High Structural Phase Purity in 2D Perovskite Films for High-Efficiency Devices
Explore the source record for details and available documents.
Thread Batching for High-performance Energy-efficient GPU Memory Design
Massive multi-threading in GPU imposes tremendous pressure on memory subsystems. Due to rapid growth in thread-level parallelism of GPU and slowly improved peak memory bandwidth, memory becomes a bottleneck of GPU’s performance and energy efficiency. In this article, we propose an integrated architectural scheme to optimize the memory accesses and therefore boost the performance and energy efficiency of GPU. First, we propose a thread batch enabled memory partitioning (TEMP) to improve GPU memory access parallelism. In particular, TEMP groups multiple thread blocks that share the same set of pages into a thread batch and applies a page coloring mechanism to bound each stream multiprocessor (SM) to the dedicated memory banks. After that, TEMP dispatches the thread batch to an SM to ensure high-parallel memory-access streaming from the different thread blocks. Second, a thread batch-aware scheduling (TBAS) scheme is introduced to improve the GPU memory access locality and to reduce the contention on memory controllers and interconnection networks. Experimental results show that the integration of TEMP and TBAS can achieve up to 10.3% performance improvement and 11.3% DRAM energy reduction across diverse GPU applications. We also evaluate the performance interference of the mixed CPU+GPU workloads when they are run on a heterogeneous system that employs our proposed schemes. Our results show that a simple solution can effectively ensure the efficient execution of both GPU and CPU applications.
Effect of virtual memory on efficient solution of two model problems
Computers with virtual memory architecture allow programs to be written as if they were small enough to be contained in memory. Two types of problems are investigated to show that this luxury can lead to quite an inefficient performance if the programmer does not interact strongly with the characteristics of the operating system when developing the program. The two problems considered are the simultaneous solutions of a large linear system of equations by Gaussian elimination and a model three-dimensional finite-difference problem. The Control Data STAR-100 computer runs are made to demonstrate the inefficiencies of programming the problems in the manner one would naturally do if the problems were indeed, small enough to be contained in memory. Program redesigns are presented which achieve large improvements in performance through changes in the computational procedure and the data base arrangement.
Memory Scalability and Efficiency Analysis of Parallel Codes
Memory scalability is an enduring problem and bottleneck that plagues many parallel codes. Parallel codes designed for High Performance Systems are typically designed over the span of several, and in some instances 10+, years. As a result, optimization practices which were appropriate for earlier systems may no longer be valid and thus require careful optimization consideration. Specifically, parallel codes whose memory footprint is a function of their scalability must be carefully considered for future exa-scale systems. In this paper we present a methodology and tool to study the memory scalability of parallel codes. Using our methodology we evaluate an application s memory footprint as a function of scalability, which we coined memory efficiency, and describe our results. In particular, using our in-house tools we can pinpoint the specific application components which contribute to the application s overall memory foot-print (application data- structures, libraries, etc.).
Lead-free epitaxial ferroelectric material integration on semiconducting (100) Nb-doped SrTiO3 for low-power non-volatile memory and efficient ultraviolet ray detection
We report lead-free ferroelectric based resistive switching non-volatile memory (NVM) devices with epitaxial (1-x)BaTiO3-xBiFeO3 (x = 0.725) (BT-BFO) film integrated on semiconducting (100) Nb (0.7%) doped SrTiO3 (Nb:STO) substrates. The piezoelectric force microscopy (PFM) measurement at room temperature demonstrated ferroelectricity in the BT-BFO thin film. PFM results also reveal the repeatable polarization inversion by poling, manifesting its potential for read-write operation in NVM devices. The electroforming-free and ferroelectric polarization coupled electrical behaviour demonstrated excellent resistive switching with high retention time, cyclic endurance and low set/reset voltages. X-ray photoelectron spectroscopy was utilized to determine the band alignment at the BT-BFO and Nb:STO heterojunction and it exhibited staggered band alignment. This heterojunction is found to behave as an efficient ultraviolet photo-detector with low rise and fall time. The architecture also demonstrates half-wave rectification under low and high input signal frequencies, where the output distortion is minimal. The results provide avenue for an electrical switch that can regulate the pixels in low or high frequency images. Combined this work paves the pathway towards designing future generation low-power ferroelectric based microelectronic devices by merging both electrical and photovoltaic properties of BT-BFO materials.
Efficient Heterogeneous Execution on Large Multicore and Accelerator Platforms: Case Study Using a Block Tridiagonal Solver
The algorithmic and implementation principles are explored in gainfully exploiting GPU accelerators in conjunction with multicore processors on high-end systems with large numbers of compute nodes, and evaluated in an implementation of a scalable block tridiagonal solver. The accelerator of each compute node is exploited in combination with multicore processors of that node in performing block-level linear algebra operations in the overall, distributed solver algorithm. Optimizations incorporated include: (1) an efficient memory mapping and synchronization interface to minimize data movement, (2) multi-process sharing of the accelerator within a node to obtain balanced load with multicore processors, and (3) an automatic memory management system to efficiently utilize accelerator memory when sub-matrices spill over the limits of device memory. Results are reported from our novel implementation that uses MAGMA and CUBLAS accelerator software systems simultaneously with ACML for multithreaded execution on processors. Overall, using 940 nVidia Tesla X2090 accelerators and 15,040 cores, the best heterogeneous execution delivers a 10.9-fold reduction in run time relative to an already efficient parallel multicore-only baseline implementation that is highly optimized with intra-node and inter-node concurrency and computation-communication overlap. Detailed quantitative results are presented to explain all critical runtime components contributing to hybrid performance.
Energy efficient photonic memory based on electrically programmable embedded III-V/Si memristors: switches and filters
Abstract Over the past few years, extensive work on optical neural networks has been investigated in hopes of achieving orders of magnitude improvement in energy efficiency and compute density via all-optical matrix-vector multiplication. However, these solutions are limited by a lack of high-speed power power-efficient phase tuners, on-chip non-volatile memory, and a proper material platform that can heterogeneously integrate all the necessary components needed onto a single chip. We address these issues by demonstrating embedded multi-layer HfO 2 /Al 2 O 3 memristors with III-V/Si photonics which facilitate non-volatile optical functionality for a variety of devices such as Mach-Zehnder Interferometers, and (de-)interleaver filters. The Mach-Zehnder optical memristor exhibits non-volatile optical phase shifts > π with ~33 dB signal extinction while consuming 0 electrical power consumption. We demonstrate 6 non-volatile states each capable of 4 Gbps modulation. (De-) interleaver filters were demonstrated to exhibit memristive non-volatile passband transformation with full set/reset states. Time duration tests were performed on all devices and indicated non-volatility up to 24 hours and beyond. We demonstrate non-volatile III-V/Si optical memristors with large electric-field driven phase shifts and reconfigurable filters with true 0 static power consumption. As a result, co-integrated photonic memristors offer a pathway for in-memory optical computing and large-scale non-volatile photonic circuits.
Parallelized Convolutional Interleaver Implementation for Efficient DDR Memory Access
Convolutional interleavers are used in many different communications systems to correct for burst errors due to atmospheric fades and scintillation. The interleaver size is related to the channel coherence time and the data rate. Small convolutional interleavers can be implemented in a field programmable gate array (FPGA) block random access memory (BRAM). However, large interleavers exceeding the size of the BRAM on the FPGA are necessary for channels with longer fades and higher data rates. Therefore, an implementation utilizing double data rate (DDR) memory external to the FPGA is necessary. Wide DDR memory data buses can make the use of DDR memory for convolutional interleavers inefficient when individual symbols are written to and read from the memory. DDR memory operational speeds can also limit the data rate of the interleaver. The Consultative Committee for Space Data Systems (CCSDS) Optical Communications High Photon Efficiency (HPE) standard utilizes a convolutional channel symbol interleaver. A previous implementation of the HPE standard utilized BRAM for the convolutional interleaver, but mission requirements for the upcoming Optical Artemis-2 Orion (O2O) communications demonstration dictate the use of an interleaver exceeding the size of the BRAM. An algorithm and method for implementing the convolutional interleaver in the FPGA with DDR memory is described in this paper.
Parallelized convolutional interleaver implementation for efficient DDR memory access
Convolutional interleavers are used in many different communications systems to correct for burst errors due to atmospheric fades and scintillation. The interleaver size is related to the channel coherence time and the data rate. Small convolutional interleavers can be implemented in a field programmable gate array (FPGA) block random access memory (BRAM). However, large interleavers exceeding the size of the BRAM on the FPGA are necessary for channels with longer fades and higher data rates. Therefore, an implementation utilizing double data rate (DDR) memory external to the FPGA is necessary. Wide DDR memory data buses can make the use of DDR memory for convolutional interleavers inefficient when individual symbols are written to and read from the memory. DDR memory operational speeds can also limit the data rate of the interleaver. The Consultative Committee for Space Data Systems (CCSDS) Optical Communications High Photon Efficiency (HPE) standard utilizes a convolutional channel symbol interleaver. A previous implementation of the HPE standard utilized BRAM for the convolutional interleaver, but mission requirements for the upcoming Optical Artemis-2 Orion (O2O) communications demonstration dictate the use of an interleaver exceeding the size of the BRAM. An algorithm and method for implementing the convolutional interleaver in the FPGA with DDR memory is described in this paper.
Expediting model-based optoacoustic reconstructions with tomographic symmetries
Purpose: Image quantification in optoacoustic tomography implies the use of accurate forward models of excitation, propagation, and detection of optoacoustic signals while inversions with high spatial resolution usually involve very large matrices, leading to unreasonably long computation times. The development of fast and memory efficient model-based approaches represents then an important challenge to advance on the quantitative and dynamic imaging capabilities of tomographic optoacoustic imaging. Methods: Herein, a method for simplification and acceleration of model-based inversions, relying on inherent symmetries present in common tomographic acquisition geometries, has been introduced. The method is showcased for the case of cylindrical symmetries by using polar image discretization of the time-domain optoacoustic forward model combined with efficient storage and inversion strategies. Results: The suggested methodology is shown to render fast and accurate model-based inversions in both numerical simulations andpost mortem small animal experiments. In case of a full-view detection scheme, the memory requirements are reduced by one order of magnitude while high-resolution reconstructions are achieved at video rate. Conclusions: By considering the rotational symmetry present in many tomographic optoacoustic imaging systems, the proposed methodology allows exploiting the advantages of model-based algorithms with feasible computational requirements and fast reconstruction times, so that its convenience and general applicability in optoacoustic imaging systems with tomographic symmetries is anticipated.
qSIEVE: Efficient qLDPC Memory via Systolic Movement in Atom Arrays
As quantum machines have scaled up in their number of qubits, significant research has turned towards increasing their fidelity with quantum error correction codes. Although promising results have been shown with the surface code, which only requires near-neighbor connections between qubits, the high qubit overhead of such local codes promises to be problematic. Consequently, recent work has explored non-local quantum LDPC (qLDPC) codes, which have good asymptotic encoding rates. Despite theoretical progress, hardware implementations of these codes have been a longstanding challenge. At the experimental level, demonstrations of movement based communication on atom arrays suggest this is a powerful new primitive to achieve non-local connectivity. Leveraging this, we present a protocol for implementing non-local qLDPC codes in hardware. Our protocol, qSIEVE, is a co-design of such codes with movement in atom arrays. qSIEVE defines a restricted family of qLDPC codes that can be implemented efficiently with systolic movement. We then quantify the utility of qSIEVE in the context of a complete fault tolerant architecture. We compare the cost of implementing benchmark programs in a standard, surface code only architecture and a mixed architecture where data is stored in qLDPC memory with qSIEVE and loaded to surface codes for computation.
Generating An Advanced Cross-section Library For HTGR Pebble Bed Depletion Calculations Using Reduced-Order Model Generation Techniques
For code development, Advanced Reactor Technologies - Gas Cooled Reactors Program (ART-GCR) rely on a collaboration with the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program, but the cross sections generation and the methodology definition is part of this program area goals. Based on previous studies in FY23, the size of microscopic cross section libraries increases rapidly with the number of tabulations, requiring significant amount of memory and drastically slowing down the Griffin calculations when evaluating cross sections via the multivariate linear interpolation approach. Rising to these challenges, this work investigates constructing Reduced-order Models (ROMs) for the multi-group microscopic cross sections to accelerate the cross section evaluation in Griffin. A database of multigroup cross sections is first collected considering all possible parameters that a designer could change for optimization. Down-selection of the ROM techniques afterward shows Deep Neural Network (DNN) as the best candidate when jointly consider memory efficiency, predictive accuracy, computational cost, scalability, flexibility and ease of implementation of the algorithms in comparison to the multidimensional interpolation. This work develops a specific interface that enables the cross section predictions using pre-trained DNN models into Griffin leveraging the existing ROM capabilities. DNNs have been trained for all isotopes for use in Griffin. Preliminary Griffin testing shows that DNNs exhibit exceptional predictive accuracy and the use of DNNs provides orders of magnitude improvement in memory efficiency compared to conventional interpolation techniques. With such ROM techniques, it holds great promise to further increase the fidelity of the Pebble Bed Reactor (PBR) simulation by increasing the number of tabulations/state variables during cross section evaluation, while maintaining the computational cost affordable in Griffin.
Electrostatically transparent graphene quantum-dot trap layers for efficient nonvolatile memory
In this study, we have demonstrated nonvolatile memory devices using graphene quantum-dots (GQDs) trap layers with indium zinc oxide (IZO) semiconductor channel. The Fermi-level of GQD was effectively modulated by tunneling electrons near the Dirac point because of limited density of states and weak electrostatic screening in monolayer graphene. As a result, large gate modulation was driven in IZO channel to achieve a subthreshold swing of 5.21 V/dec (300 nm SiO{sub 2} gate insulator), while Au quantum-dots memory shows 15.52 V/dec because of strong electrostatic screening in metal quantum-dots. Together, discrete charge traps of GQDs enable stable performance in the endurance test beyond 800 cycles of programming and erasing. Our study suggests the exciting potential of GQD trap layers to be used for a highly promising material in non-volatile memory devices.
Characterization of Memory Devices for Energy Efficient Analog In-Memory Neural Computing at the Edge.
Abstract not provided.
What Scientific Applications can Benefit from Hardware Transactional Memory?
Achieving efficient and correct synchronization of multiple threads is a difficult and error-prone task at small scale and, as we march towards extreme scale computing, will be even more challenging when the resulting application is supposed to utilize millions of cores efficiently. Transactional Memory (TM) is a promising technique to ease the burden on the programmer, but only recently has become available on commercial hardware in the new Blue Gene/Q system and hence the real benefit for realistic applications has not been studied, yet. This paper presents the first performance results of TM embedded into OpenMP on a prototype system of BG/Q and characterizes code properties that will likely lead to benefits when augmented with TM primitives. We first, study the influence of thread count, environment variables and memory layout on TM performance and identify code properties that will yield performance gains with TM. Second, we evaluate the combination of OpenMP with multiple synchronization primitives on top of MPI to determine suitable task to thread ratios per node. Finally, we condense our findings into a set of best practices. These are applied to a Monte Carlo Benchmark and a Smoothed Particle Hydrodynamics method. In both cases an optimized TM version, executed with 64 threads on one node, outperforms a simple TM implementation. MCB with optimized TM yields a speedup of 27.45 over baseline.
On-the-fly Doppler broadening of unresolved resonance region cross sections
In this paper, two methods for computing temperature-dependent unresolved resonance region cross sections on-the-fly within continuous-energy Monte Carlo neutron transport simulations are presented. The first method calculates Doppler broadened cross sections directly from zero-temperature average resonance parameters. In a simulation, at each event that requires cross section values, a realization of unresolved resonance parameters is generated about the desired energy and temperature-dependent single-level Breit-Wigner resonance cross sections are computed directly via the analytical Ψ-x Doppler integrals. The second method relies on the generation of equiprobable cross section magnitude bands on an energy-temperature mesh. Within a simulation, the bands are sampled and interpolated in energy and temperature to obtain cross section values on-the-fly. Both of the methods, as well as their underlying calculation procedures, are verified numerically in extensive code-to-code comparisons. Energy-dependent pointwise cross sections calculated with the newly-implemented procedures are shown to be in excellent agreement with those calculated by a widely-used nuclear data processing code. Relative differences at or below 0.1% are observed. Integral criticality benchmark results computed with the proposed methods are shown to reproduce those computed with a state-of-the-art processed nuclear data library very well. In simulations of fast spectrum systems which are highly-sensitive to the representation of cross section data in the unresolved region, k-eigenvalue and neutron flux spectra differences of <10 pcm and <1.0% are observed, respectively. The direct method is demonstrated to be well-suited to the calculation of reference solutions — against which results obtained with a discretized representation may be assessed — as a result of its treatment of the energy, temperature, and cross section magnitude variables as continuous. Also, because there is no pre-processed data to store (only temperature-independent average resonance parameters) the direct method is very memory-efficient. Typically, only a few kB of memory are needed to store all required unresolved region data for a single nuclide. However, depending on the details of a particular simulation, performing URR cross section calculations on-the-fly can significantly increase simulation times. Alternatively, the method of interpolating equiprobable probability bands is demonstrated to produce results that are as accurate as the direct reference solutions, to within arbitrary precision, with high computational efficiency in terms of memory requirements and simulation time. Analyses of a fast spectrum system show that interpolation on a coarse energy-temperature mesh can be used to reproduce reference k-eigenvalue results obtained with cross sections calculated continuously in energy and directly at an exact temperature to within <10 pcm. Probability band data on a mesh encompassing the range of temperatures relevant to reactor analysis usually require around 100 kB of memory per nuclide. Finally, relative to the case in which probability table data generated at a single, desired temperature are used, minor increases in simulation times are observed when probability band interpolation is employed.
Using SciDB to Support Photon Science Data Analysis
Array data analytic systems like SciDB hold great potential to accelerate processing data from SLAC's Linac Coherent Light Source and other experiments. SciDB is unique in its ability to integrate storage and processing of array data efficiently, providing both space-efficient storage and out-of-memory efficient parallel array processing. We describe a recent effort to leverage SciDB to store and process LCLS data. The work includes development of software to import data into SciDB, subsequent benchmarks, and interactive manipulation of data in SciDB.