SEARCH · Search NASA

Results for “bandwidth optimization”

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.

SourceResearch connectionFrom yearThrough year

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

At least 271 records · Page 15

All sources NASA NTRS DOE OSTI

DOE OSTI2023

System and method for high dynamic range waveform digitization

Diverse applications from particle physics experiments to lidar are driving cost and current reduction in giga-hertz sampling rate high-resolution data conversion. Multiple imagers captures a single pixel of data and require processing at very high speed. High-bandwidth high-rate signal sampling, analog-to-digital conversion, and transfer of large amounts of data to a digital data acquisition block are required in such systems. Dynamic range, power consumption, and transfer of high-speed, high-bit width data are key implementation challenges. Data acquisition architectures optimized for specific requirements of such systems may facilitate system implementation and reduce overall system cost.

Mostafanezhad, Isar↗

DOE OSTI2020

Population Count on Intel® CPU, GPU, and FPGA

Population count is a primitive used in many applications. Commodity processors have dedicated instructions for achieving high-performance population count. Motivated by the productivity of high-level synthesis and the importance of population count, in this paper we investigated the OpenCL implementations of population count algorithms, and evaluated their performance and resource utilizations on an FPGA. Based on the results, we select the most efficient implementation. Then we derived a reduction pattern from a representative application of population count. We parallelized the reduction with atomic functions, and optimized it with vectorized memory accesses, tree reduction, and compute-unit duplication. We evaluated the performance of the reduction kernel on an Intel (R) Xeon (R) CPU and an Intel (R) IrisT Pro integrated GPU, and an FPGA card that features an Intel (R) Arria (R) 10 FPGA. When DRAM memory bandwidth is comparable on the three computing platforms, the FPGA can achieve the highest kernel performance for large workload. On the other hand, we described performance bottlenecks on the FPGA. To make FPGAs more competitive in raw performance compared to high-performant CPU and GPU platforms, it is important to increase external memory bandwidth, minimize data movement between a host and a device, and reduce OpenCL runtime overhead on an FPGA.

heterogeneous computing↗

DOE OSTI2017

Higher-order triangular spectral element method with optimized cubature points for seismic wavefield modeling

The mass-lumped method avoids the cost of inverting the mass matrix and simultaneously maintains spatial accuracy by adopting additional interior integration points, known as cubature points. To date, such points are only known analytically in tensor domains, such as quadrilateral or hexahedral elements. Thus, the diagonal-mass-matrix spectral element method (SEM) in non-tensor domains always relies on numerically computed interpolation points or quadrature points. However, only the cubature points for degrees 1 to 6 are known, which is the reason that we have developed a p-norm-based optimization algorithm to obtain higher-order cubature points. In this way, we obtain and tabulate new cubature points with all positive integration weights for degrees 7 to 9. The dispersion analysis illustrates that the dispersion relation determined from the new optimized cubature points is comparable to that of the mass and stiffness matrices obtained by exact integration. Simultaneously, the Lebesgue constant for the new optimized cubature points indicates its surprisingly good interpolation properties. As a result, such points provide both good interpolation properties and integration accuracy. The Courant–Friedrichs–Lewy (CFL) numbers are tabulated for the conventional Fekete-based triangular spectral element (TSEM), the TSEM with exact integration, and the optimized cubature-based TSEM (OTSEM). A complementary study demonstrates the spectral convergence of the OTSEM. A numerical example conducted on a half-space model demonstrates that the OTSEM improves the accuracy by approximately one order of magnitude compared to the conventional Fekete-based TSEM. In particular, the accuracy of the 7th-order OTSEM is even higher than that of the 14th-order Fekete-based TSEM. Furthermore, the OTSEM produces a result that can compete in accuracy with the quadrilateral SEM (QSEM). The high accuracy of the OTSEM is also tested with a non-flat topography model. In terms of computational efficiency, the OTSEM is more efficient than the Fekete-based TSEM, although it is slightly costlier than the QSEM when a comparable numerical accuracy is required. - Highlights: • Higher-order cubature points for degrees 7 to 9 are developed. • The effects of quadrature rule on the mass and stiffness matrices has been conducted. • The cubature points have always positive integration weights. • Freeing from the inversion of a wide bandwidth mass matrix. • The accuracy of the TSEM has been improved in about one order of magnitude.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2022

Design considerations for gallium arsenide pulse compression photoconductive switch

In this paper, we present the physics and design-space exploration of a novel pulse compression photoconductive switch (PCPS) using semi-insulating gallium arsenide (GaAs) operating in the negative differential mobility (NDM) regime of electron transport. We systematically quantify the relationship between the PCPS performance and various design options, including contact separation, laser energy and placement, and trap dynamics. Specifically, we report the full-width at half-maximum and the peak output current generated by the PCPS as a function of applied electrical and optical bias. We discuss the optimal spacing between the electrodes and the distance of the laser spot to the anode to achieve higher electron confinement and superior radio-frequency (RF) metrics. Reducing the laser energy is important to prevent the appearance of secondary peaks due to diffusive transport, but there exists a trade-off between the bandwidth and the maximum current of the PCPS. We also compare the PCPS response with and without trap dynamics and find that the electrostatic screening from the trap-induced space charge is time-independent when the trapping time constant is set larger than the recombination lifetime. Overall, trap dynamics are detrimental to performance, unless the compensation doping scheme to achieve semi-insulating GaAs is carefully selected. Results presented in this paper can be used by experimentalists to fine-tune the PCPS design parameters to meet the specifications of various RF applications. Moreover, our results will provide a strong theoretical basis to the measurements of PCPS devices using GaAs and other NDM materials under investigation.

42 ENGINEERING↗

DOE OSTI2013

Grid artifact reduction for direct digital radiography detectors based on rotated stationary grids with homomorphic filtering

Purpose: Grid artifacts are caused when using the antiscatter grid in obtaining digital x-ray images. In this paper, research on grid artifact reduction techniques is conducted especially for the direct detectors, which are based on amorphous selenium. Methods: In order to analyze and reduce the grid artifacts, the authors consider a multiplicative grid image model and propose a homomorphic filtering technique. For minimal damage due to filters, which are used to suppress the grid artifacts, rotated grids with respect to the sampling direction are employed, and min-max optimization problems for searching optimal grid frequencies and angles for given sampling frequencies are established. The authors then propose algorithms for the grid artifact reduction based on the band-stop filters as well as low-pass filters. Results: The proposed algorithms are experimentally tested for digital x-ray images, which are obtained from direct detectors with the rotated grids, and are compared with other algorithms. It is shown that the proposed algorithms can successfully reduce the grid artifacts for direct detectors. Conclusions: By employing the homomorphic filtering technique, the authors can considerably suppress the strong grid artifacts with relatively narrow-bandwidth filters compared to the normal filtering case. Using rotated grids also significantly reduces the ringing artifact. Furthermore, for specific grid frequencies and angles, the authors can use simple homomorphic low-pass filters in the spatial domain, and thus alleviate the grid artifacts with very low implementation complexity.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2017

Packet Capture Solutions: PcapDB Benchmark for High-Bandwidth Capture, Storage, and Searching

PcapDB stands alone when looking at the overall field of competitors, from the cost-effective COTS hardware, to the efficient utilization of disk space that enables a longer packet history. A scalable, 100GbE-enabled system that indexes every packet and indexes flow data without complicated load-balancing requirements. The Transport Layer search and indexing approach led to patent-pending flow indexing technology, providing a specialized database system specifically optimized around providing fast flow searches. While there are a plethora of options in network packet capture, there are very few that are able to effectively manage capture rates of more than 10 Gb/s, distributed capture and querying, and a responsive user interface. By far, the primary competitor in the market place is Endace and DeepSee; in addition to meeting the technical requirements we set out in this document, they provide technical support and a fully 'appliance like' system. In terms of cost, however, our experience has been that the yearly maintenance charges alone outstrip the entire hardware cost of solutions like PcapDB. Investment in cyber security research and development is a large part of what has enabled us to build the base of knowlegable workers needed to defend government resources in the rapidly evolving cyber security landscape. We believe projects like Bro, WireCap, and Farm do more than just fill temporary gaps in our capabilities. They give allow us to build the firm foundation needed to tackle the next generation of cyber challenges. PcapDB was built with loftier ambitions than simply solving the packet capture of a single lab site, but instead to provide a robust, scaleable packet capture solution to the DOE complex and beyond.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

DOE OSTI2024

The ECP SICM project: Managing complex memory hierarchies for exascale applications

The Exascale Computing Project (ECP)’s Simplified Interface to Complex Memories (SICM) effort focuses on developing universal interfaces for discovering, managing, and sharing data across complex memory hierarchies. These facilitate the exploitation of emerging memory technologies and support precise control over their various trade-offs such as high-bandwidth versus low-latency, persistent versus ephemeral, high-capacity versus low-capacity, and near-CPU versus near-GPU. SICM comprises three interrelated components: a low-level interface, a high-level interface, and a persistent-heap interface. The low-level SICM interface is intended for system and run-time developers as well as expert application developers who prefer full control of the memory objects used within their application. The high-level SICM interface builds upon the low-level interface, employing application-level profiling and analysis to optimize data management for complex memory hierarchies. The persistent-heap interface provides applications with a persistent memory allocator that can allocate custom C++ data structures in both block-storage and byte-addressable persistent memories.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

SMALE: Enhancing Scalability of Machine Learning Algorithms on Extreme-Scale Computing Platforms

Deployment and execution of machine learning tasks on extreme-scale computing platforms face several significant technical challenges: 1) High computing cost incurred by dense networks – The computing workload of deep networks with densely-connected topology increases rapidly with the network size, imposing a non-scalable computing model of extreme-scale computing platforms; 2) Non-optimized workload distribution – Many advanced deep learning algorithms, e.g., sparsification and irregular net-work topology, produce very unbalanced workload distribution on extreme-scale computing platforms. The computation efficiency is greatly hindered by the incurred data and computation redundancies as well as long tails of the node with extensive workload; 3) Constraints in data movement and I/O bottle-neck – Inter-node data movement in extreme-scale computing platforms are associated with high energy and latency costs, and subject to the constraints of I/O bandwidth; and 4) Generalization of algorithm realization and acceleration on computing platforms – The large varieties of machine learning algorithms and structures of extreme-scale computing platforms make the derivation of a generalized algorithm realization and acceleration method very challenging, which, however, is the requirement by domain scientists and interested users. We call the above challenges Smale’s Problems in Machine Learning and Understanding for High-Performance Computing Scientific Discovery. The objective of our three-year research project is to develop a holistic innovation set at structure, assembly, and acceleration layers of machine learning algorithms to address the above challenges in algorithm deployment and execution. Three tasks are particularly performed, including: At the algorithm structure level, we investigate the techniques that can structurally sparsify on the topology of deep networks for computing workload reduction. We also study clustering and pruning techniques that can optimize the workload distributions over the extreme-scale computing platforms; At the algorithm assembly level, we derive a unified learning framework for unsupervised transfer learning and dynamic growing capabilities. Novel training methods are also exploited to enhance the training efficiency of the proposed framework; At the algorithm acceleration level, we will develop a series of techniques that can accelerate the computation of sparse matrix operations, which are one of the core executions in deep learning and optimize memory access of the concerned platforms. Our proposed techniques attack the fundamental problems in machine learning algorithms running on extreme-scale computing platforms by vertically integrating the solutions at three closely entangled layers, paving the long-term scaling path of machine learning applications under DOE context. Three tasks corresponding to the above respective research orientations are performed during the three-year project period with our collaborators at ORNL. The outcome of the proposed project is anticipated to form a holistic solution set of novel algorithms and network topologies, efficient training techniques, and fast acceleration methods to promote the computing scalability of the machine learning applications of particular interest to DOE.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2016

SIFT-based dense pixel tracking on 0.35 T cine-MR images acquired during image-guided radiation therapy with application to gating optimization

Purpose: To first demonstrate the viability of applying an image processing technique for tracking regions on low-contrast cine-MR images acquired during image-guided radiation therapy, and then outline a scheme that uses tracking data for optimizing gating results in a patient-specific manner. Methods: A first-generation MR-IGRT system—treating patients since January 2014—integrates a 0.35 T MR scanner into an annular gantry consisting of three independent Co-60 sources. Obtaining adequate frame rates for capturing relevant patient motion across large fields-of-view currently requires coarse in-plane spatial resolution. This study initially (1) investigate the feasibility of rapidly tracking dense pixel correspondences across single, sagittal plane images (with both moderate signal-to-noise and spatial resolution) using a matching objective for highly descriptive vectors called scale-invariant feature transform (SIFT) descriptors associated to all pixels that describe intensity gradients in local regions around each pixel. To more accurately track features, (2) harmonic analysis was then applied to all pixel trajectories within a region-of-interest across a short training period. In particular, the procedure adjusts the motion of outlying trajectories whose relative spectral power within a frequency bandwidth consistent with respiration (or another form of periodic motion) does not exceed a threshold value that is manually specified following the training period. To evaluate the tracking reliability after applying this correction, conventional metrics—including Dice similarity coefficients (DSCs), mean tracking errors (MTEs), and Hausdorff distances (HD)—were used to compare target segmentations obtained via tracking to manually delineated segmentations. Upon confirming the viability of this descriptor-based procedure for reliably tracking features, the study (3) outlines a scheme for optimizing gating parameters—including relative target position and a tolerable margin about this position—derived from a probability density function that is constructed using tracking results obtained just prior to treatment. Results: The feasibility of applying the matching objective for SIFT descriptors toward pixel-by-pixel tracking on cine-MR acquisitions was first retrospectively demonstrated for 19 treatments (spanning various sites). Both with and without motion correction based on harmonic analysis, sub-pixel MTEs were obtained. A mean DSC value spanning all patients of 0.916 ± 0.001 was obtained without motion correction, with DSC values exceeding 0.85 for all patients considered. While most patients show accurate tracking without motion correction, harmonic analysis does yield substantial gain in accuracy (defined using HDs) for three particularly challenging subjects. An application of tracking toward a gating optimization procedure was then demonstrated that should allow a physician to balance beam-on time and tissue sparing in a patient-specific manner by tuning several intuitive parameters. Conclusions: Tracking results show high fidelity in assessing intrafractional motion observed on cine-MR acquisitions. Incorporating harmonic analysis during a training period improves the robustness of the tracking for challenging targets. The concomitant gating optimization procedure should allow for physicians to quantitatively assess gating effectiveness quickly just prior to treatment in a patient-specific manner.

61 RADIATION PROTECTION AND DOSIMETRY↗

DOE OSTI2018

Adaptive Runtime Features For Distributed Graph Algorithms

Performance of distributed graph algorithms can benefit greatly by forming rapport between algorithmic abstraction and the underlying runtime system that is responsible for scheduling work and exchanging messages. However, due to their dynamic and irregular nature of computation, distributed graph algorithms written in different programming models impose varying degree of workload pressure on the runtime. To cope with such vastly different workload characteristics, a runtime has to make several trade-offs. One such trade-off arises, for example, when the runtime scheduler has to choose among alternatives such as whether to execute algorithmic work, or progress the network by probing network buffers, or throttle sending messages (termed flow control). This tradeoff decides between optimizing the throughput of a runtime scheduler by increasing the rate of execution of algorithmic work, and reducing the latency of the network messages. Another trade-off exists when a decision has to be made about when to send aggregated messages in buffers (message coalescing). This decision chooses between trading off latency for network bandwidth and vice versa. At any instant, such trade-offs emphasize either on improving the quantity of work being executed (by maximizing the scheduler throughput) or on improving the quality of work (by prioritizing better work). However, encoding static policies for different runtime features (such as flow control, coalescing) can prevent graph algorithms from achieving their full potential, thus can undermine the actual performance of a distributed graph algorithm . In this paper, we investigate runtime support for distributed graph algorithms in the context of two paradigms: variants of wellknown Bulk-Synchronous Parallel model and asynchronous programming model. We explore generic runtime features such as message coalescing (aggregation) and flow control and show that execution policies of these features need to be adjusted over time to make a positive impact on the execution time of a distributed graph algorithm. Since synchronous and asynchronous graph algorithms have different workload characteristics, not all of such runtime features may be good candidates for adaptation. Each of these algorithmic paradigms may require different set of features to be adapted over time. We demonstrate which set of feature(s) can be useful in each case to achieve the right balance of work in the runtime layer. Existing implementation of different graph algorithms can benefit from adapting dynamic policies in the underlying runtime.

Firoz, Jesun S.↗

DOE OSTI2023

Design and simulation of multilayer coatings for a multi-channel Wolter-like x-ray imager with large field of view and high resolution

X-ray diagnostics are key instruments for understanding the physics behind inertial confinement fusion experiments. We report on the multilayer design optimization for the Toroidal X-ray Imager (TXI), a hard x-rays microscope instrument designed by Commissariat à l’énergie atomique (CEA) and Laboratoire Charles Fabry (LCF) to be installed on the National Ignition Facility. TXI includes six channels designed for three different energy bands centered on 8.7, 13, and 17.5 keV. Each channel is made up of two toroidal mirrors arranged in a Wolter-like configuration. The required field of view is 800 × 400 µm2, and the resolution should be better than 5 µm. In addition, we seek to estimate the spatial distribution of the temperature, which requires no spectral overlap of the different energy bands and a good spectral homogeneity of the image produced. The development of the multilayer coatings was performed in a two-step method. First, the coatings were optimized to obtain proper energy bands. Then, an x-ray tracing code was used to calculate the integrated optical response of each channel and adjust the response of the mirror to fulfill the requirements. To fulfill all the specifications, we propose an original design using a combination of two aperiodic coatings, one with a narrow bandwidth and the other one with a larger bandwidth.

hard x-rays↗

DOE OSTI2024

The 140 GHz notch filter development for millimeter-wave diagnostics protection on the stellarator Wendelstein 7-X

The notch filter plays a crucial role as a protective component in microwave diagnostics, primarily by addressing issues related to catastrophic interference. Designed for millimeter-wave diagnostics on the stellarator Wendelstein 7-X (W7-X), a WR-6 waveguide-based notch filter has been successfully developed to effectively isolate leakage from auxiliary heating gyrotrons operating at 140 GHz. The filter incorporates cylindrical cavities resonating at 140 GHz for the TE11p mode, with coupling structures that are designed and optimized for high-efficiency coupling. This configuration simplifies fabrication, thereby ensuring high-yield production. Experimental fabrication and in-house characterization confirm the notch filter's exceptional performance, with over 60 dB rejection in the vicinity of 140 GHz and low insertion loss (< 2 dB) above and below the notch frequency across a broad frequency bandwidth (121–138 GHz, 142–163 GHz). Furthermore, the utilization of this high-frequency structure fabrication technology can be applied to millimeter-wave diagnostics on other machines. In addition to the design elements of the notch filter, this paper also provides a detailed discussion of the fabrication process and methodology.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2020

A High-Throughput Solver for Marginalized Graph Kernels on GPU

Here, we present the design and optimization of a solver for efficient and high-throughput computation of the marginalized graph kernel on General Purpose GPUs. The graph kernel is computed using the conjugate gradient method to solve a generalized Laplacian of the tensor product between a pair of graphs. To cope with the large gap between the instruction throughput and the memory bandwidth of the GPUs, our solver forms the graph tensor product on-the-fly without storing it in memory. This is achieved by using threads in a warp cooperatively to stream the adjacency and edge label matrices of individual graphs by small square matrix blocks called tiles, which are then staged in registers and the shared memory for later reuse. Warps across a thread block can further share tiles via the shared memory to increase data reuse. We exploit the sparsity of the graphs hierarchically by storing only non-empty tiles using a coordinate format and nonzero elements within each tile using bitmaps. We propose a new partition-based reordering algorithm for aggregating nonzero elements of the graphs into fewer but denser tiles to further exploit sparsity. We carry out extensive theoretical analyses on the graph tensor product primitives for tiles of various density and evaluate their performance on synthetic and real-world datasets. Our solver delivers three to four orders of magnitude speedup over existing CPU-based solvers such as GraKeL and GraphKernels. The capability of the solver enables kernel-based learning tasks at unprecedented scales.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2012

Efficient Databases for MPC Microdata (Final Report)

The purpose of this grant was to develop the theory and practice of high-performance databases for massive streamed datasets. Over the last three years, we have developed fast indexing technology, that is, technology for rapidly ingesting data and storing that data so that it can be efficiently queried and analyzed. During this project we developed the technology so that high-bandwidth data streams can be indexed and queried efficiently. Our technology has been proven to work data sets composed of tens of billions of rows when the data streams arrives at over 40,000 rows per second. We achieved these numbers even on a single disk driven by two cores. Our work comprised (1) new write-optimized data structures with better asymptotic complexity than traditional structures, (2) implementation, and (3) benchmarking. We furthermore developed a prototype of TokuFS, a middleware layer that can handle microdata I/O packaged up in an MPI-IO abstraction.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Noise limits for dc SQUID readout of high- Q resonators below 300 MHz

We present the limits on noise for the readout of cryogenic high-Q resonators using dc Superconducting Quantum Interference Devices (SQUIDs) below 300 MHz. This analysis uses realized first-stage SQUIDs (previously published), whose performance is well described by Tesche–Clarke (TC) theory, coupled directly to the resonators. We also present data from a prototype second-stage dc SQUID array designed to couple to this first-stage SQUID as a follow-on amplifier with high system bandwidth. This analysis is the first full consideration of dc SQUID noise performance referred to a high-Q resonator over this frequency range and is presented relative to the standard quantum limit. We include imprecision, backaction, and backaction–imprecision noise correlations from TC theory, the noise contributed by the second-stage SQUIDs, wiring, and preamplifiers, and optimizations for both on-resonance measurements and off-resonance scan sensitivity. This architecture has modern relevance due to the increased interest in axion searches and the requirements of the DMRadio-m 3 axion search, which uses dc SQUIDs in this frequency range.

Bosons↗

DOE OSTI2026

Towards a self-driving trigger at the LHC: adaptive response in real time

Real-time data filtering and selection—or trigger—systems at high-throughput scientific facilities such as the experiments at the Large Hadron Collider must process extremely high-rate data streams under stringent bandwidth, latency, and storage constraints. Yet these systems are typically designed as static, hand-tuned menus of selection criteria grounded in prior knowledge and simulation. In this work, we further explore the concept of a self-driving trigger, an autonomous data-filtering framework that reallocates resources and adjusts thresholds dynamically in real-time to optimize signal efficiency, rate stability, and computational cost as instrumentation and environmental conditions evolve. We introduce a benchmark ecosystem to emulate realistic collider scenarios and demonstrate real-time optimization of a menu including canonical energy sum triggers as well as modern anomaly-detection algorithms that target non-standard event topologies using machine learning. Using simulated data streams and publicly available collision data from the Compact Muon Solenoid experiment, we demonstrate the capability to dynamically and automatically optimize trigger performance under specific cost objectives without manual retuning. Our adaptive strategy shifts trigger design from static menus with heuristic tuning to intelligent, automated, data-driven control, unlocking greater flexibility and discovery potential in future high-energy physics analyses.

Emami, Shaghayegh [Michigan U.] (ORCID:00090007589↗

DOE OSTI2023

Software-Hardware Co-design of Heterogeneous SmartNIC System for Recommendation Models Inference and Training

Deep Learning Recommendation Models (DLRMs) are critical applications in various domains and have evolved as one of the single largest machine learning applications. Trillions of DLRM parameters exceed the on-chip memory capacity of GPUs. Large-scale multi-node systems are required for distributed DLRM inference and training, which suffer from the all-to-all communication bottleneck, mainly limiting the scalability of ever-growing DLRMs. In recent years, SmartNICs have evolved with coupled computation and communication capabilities providing opportunities for a powerful heterogeneous device in the system. However, there isn't such a distributed system that fully leverages the abundant smartNIC resources that resolve the scalability issue of DLRMs. In this work, we proposed a software-hardware co-design of a heterogeneous smartNIC system that resolves the communication bottleneck of distributed DLRMs, mitigates the memory bandwidth pressure, and improves computation efficiency. We provide a set of smartNIC designs of cache systems (including local cache and remote cache) and smartNIC computation kernels which reduce data movement, relieve memory lookup intensity, and improve the GPU's computation efficiency. In addition, we propose a graph algorithm that improves the data locality of queries within batches which optimizes the overall system performance with higher data reuse. Our evaluation shows that our system achieves 2.1x latency speedup for inference and 1.6x throughput speedup for training.

Guo, Anqi↗

DOE OSTI2023

Better Than Worst-Case Decoding for Quantum Error Correction

The overheads of classical decoding for quantum error correction in cryogenic quantum systems grow rapidly with the number of logical qubits and their correction code distance. Decoding at room temperature is bottlenecked by refrigerator I/O bandwidth while cryogenic on-chip decoding is limited by area/power/thermal budget. To overcome these overheads, we are motivated by the observation that in the common case (over 90% of the time), error correction 'syndromes' are fairly trivial with high redundancy / sparsity, since the error correction codes are over-provisioned to be able to correct for uncommon worst-case complex scenarios (to ensure substantially low logical error rates). If suitably exploited, these trivial scenarios can be handled with insignificant overhead, thereby alleviating any bottlenecks towards handling the worst-case scenarios by state-of-the-art means. Here, we propose Better Than Worst-Case Decoding for Quantum Error Correction, targeting cryogenic quantum systems and Surface Code, consisting of: On-chip Clique Decoder: An extremely lightweight decoder for correcting trivial common-case errors, designed for the cryogenic domain. The decoder is implemented and evaluated for SFQ logic. Statistical Off-chip Bandwidth Allocation: A statistical confidence-based technique for allocation of off-chip decoding bandwidth, to efficiently handle the rare complex decodes that are not covered by the Clique Decoder. Decode-Overflow Execution Stalling: A method to stall circuit execution, for the worst-case scenarios in which the provisioned off-chip bandwidth is insufficient to complete all requested off-chip decodes. In all, BTWC decoding achieves 70-99+% off-chip bandwidth elimination across a range of logical and physical error rates, without significantly sacrificing the accuracy of a state-of-the-art off-chip decoder. Further, it achieves 10-1000x bandwidth reduction over prior bandwidth reduction techniques, as well as 15-37x resource overhead reduction compared to prior on-chip decoding.

cryogenic optimization↗