Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “FPGA implementation”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Real-Time Inference For MI/RR Deblending

The Fermilab Main Injector (MI) and Recycler Ring (RR) share a common beam loss monitor (BLM) system, making loss events difficult to attribute to their source machine when beam is present in both simultaneously. The Real-time Edge AI for Distributed Systems (READS) project addresses this by deblending BLM readings in real time using machine learning (ML). The current FPGA based implementation meets the sub-3 ms latency requirement but carries a resource intensive hls4ml development cycle, motivating exploration of GPU based deployment. This paper characterizes inference latency on an NVIDIA Jetson Orin Nano and introduces a packet organization scheme for assembling synchronized event frames from seven distributed BLM DAQ streams. Using a Python based DAQ simulation with injected timing jitter in place of unavailable live beam data, the pipeline achieved an average end to end latency of 0.456 ms (σ = 0.122 ms) across 167,000 test frames, comfortably meeting the timing constraint. Early outliers were attributed to TensorRT warm-up rather than steady state limitations, suggesting GPU based inference is a viable alternative to the existing FPGA implementation.

Yu, Kellen [Cornell U.]↗
DOE OSTI2026

Exploration of Real Time Inference for MI-RR Deblending on GPU/TPU Systems

The Fermilab Main Injector (MI) and Recycler Ring (RR) share a common beam loss monitor (BLM) system, making loss events difficult to attribute to their source machine when beam is present in both simultaneously. The Real-time Edge AI for Distributed Systems (READS) project addresses this by deblending BLM readings in real time using machine learning (ML). The current FPGA based implementation meets the sub-3 ms latency requirement but carries a resource intensive hls4ml development cycle, motivating exploration of GPU based deployment. This paper characterizes inference latency on an NVIDIA Jetson Orin Nano and introduces a packet organization scheme for assembling synchronized event frames from seven distributed BLM DAQ streams. Using a Python based DAQ simulation with injected timing jitter in place of unavailable live beam data, the pipeline achieved an average end to end latency of 0.456 ms (σ = 0.122 ms) across 167,000 test frames, comfortably meeting the timing constraint. Early outliers were attributed to TensorRT warm-up rather than steady state limitations, suggesting GPU based inference is a viable alternative to the existing FPGA implementation.

Yu, Kellen [Fermilab; Cornell U.]↗
DOE OSTI2020

An 8.8 ps RMS Resolution Time-To-Digital Converter Implemented in a 60 nm FPGA with Real-Time Temperature Correction

This paper presented a non-uniform multiphase (NUMP) time-to-digital converter (TDC) implemented in a field-programmable gate array (FPGA) with real-time automatic temperature compensation. NUMP-TDC is a novel, low-cost, high-performance TDC that has achieved an excellent performance in Altera Cyclone V FPGA. The root mean square (RMS) for the intrinsic timing resolution was 2.3 ps. However, the propagation delays in the delay chain of some FPGAs (for example, the Altera Cyclone 10 LP) vary significantly as the temperature changes. Thus, the timing performances of NUMP-TDCs implemented in those FPGAs are significantly impacted by temperature fluctuations. In this study, a simple method was developed to monitor variations in propagation delays using two registers deployed at both ends of the delay chain and compensate for changes in propagation delay using a look-up table (LUT). When the variations exceeded a certain threshold, the LUT for the delay correction was updated, and a bin-by-bin correction was launched. Using this correction approach, a resolution of 8.8 ps RMS over a wide temperature range (5 °C to 80 °C) had been achieved in a NUMP-TDC implemented in a Cyclone 10 LP FPGA.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

A demonstrator for a real-time AI-FPGA-based triggering system for sPHENIX at RHIC

The RHIC interaction rate at sPHENIX will reach around 3 MHz in pp collisions and requires the detector readout to reject events by a factor of over 200 to fit the DAQ bandwidth of 15 kHz. Some critical measurements, such as heavy flavor production in pp collisions, often require the analysis of particles produced at low momentum. This prohibits adopting the traditional approach, where data rates are reduced through triggering on rare high momentum probes. We explore a new approach based on real-time AI technology, adopt an FPGA-based implementation using a custom designed FELIX-712 board with the Xilinx Kintex Ultrascale FPGA, and deploy the system in the detector readout electronics loop for real-time trigger decision.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

A demonstrator for a real-time AI-FPGA-based triggering system for sPHENIX at RHIC

The RHIC interaction rate at sPHENIX will reach around 3 MHz in pp collisions and requires the detector readout to reject events by a factor of over 200 to fit the DAQ bandwidth of 15 kHz. Some critical measurements, such as heavy flavor production in pp collisions, often require the analysis of particles produced at low momentum. This prohibits adopting the traditional approach, where data rates are reduced through triggering on rare high momentum probes. We explore a new approach based on real-time AI technology, adopt an FPGA-based implementation using a custom designed FELIX-712 board with the Xilinx Kintex Ultrascale FPGA, and deploy the system in the detector readout electronics loop for real-time trigger decision.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

In-pixel automatic threshold calibration for the CMS Endcap Timing Layer readout chip

We present the implementation and verification of an in-pixel automatic threshold calibration circuit for the CMS Endcap Timing Layer (ETL) in the High-Luminosity LHC upgrade. The discriminator threshold of the ETL readout chip (ETROC) needs to be calibrated regularly to mitigate the circuit baseline change. Traditional methods need a lot of communication through a slow control system hence are time-consuming. This paper describes an in-pixel automatic scheme with improvements in operating time and usability. In this scheme, a sample-accumulation circuit is used to measure the average discriminator output. A binary successive approximation and linear combination scan are applied to find the equivalent baseline. The actual calibration procedure has been first implemented in FPGA firmware and tested with the ETROC front-end prototype chip (ETROC0). The calibration circuit has been implemented with Triple Modular Redundancy (TMR) and verified with Single Event Effects (SEEs) simulation. A complete calibration process lasts 35 ms with a 40 MHz clock. In the worst case, the dynamic and static power consumption are estimated to be 300 uW and 10.4 uW, respectively. The circuit design, implemented in a 65 CMOS technology, will be integrated into ETROC2, the next iteration of the ETROC with a 16 x16 pixel matrix.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

Optimizing FPGA-based Accelerator Design for Large-Scale Molecular Similarity Search (Special Session Paper)

Molecular similarity search has been widely used in drug discovery to rapidly identify structurally similar compounds from large molecular databases. With the increasing size of chemical libraries, there is growing interest in the efficient ac- celeration of large-scale similarity search. Existing works mainly focus on CPU and GPU to accelerate the computation of Tatimoto coefficient in measuring the pairwise similarity between different molecular fingerprints. In this paper, we propose and optimize an FPGA-based accelerator design on exhaustive and approximate search algorithms. On exhaustive search using BitBound & fold- ing, we analyze the similarity cutoff and folding level relationship with search speedup and accuracy, and propose a scalable on- the-fly query engine on FPGAs to reduce the resource utilization and pipeline interval. We achieve a 450 million compounds-per- second processing throughput for a single query engine. On approximate search using hierarchical navigable small world (HNSW), a popular algorithm with high recall and query speed, we propose an FPGA-based graph traversal engine to utilize high throughput register array based priority queue and fine- grained distance calculation engine to increase the processing capability. Experimental results show that the proposed FPGA- based HNSW implementation achieves a 35× speedup than existing works on CPU. To the best of our knowledge, our FPGA- based implementation is the first attempt to accelerate molecular similarity search on FPGA and has the highest performance among existing approaches.

Peng, Hongwu↗
DOE OSTI2021

The Event Detection System in the NEXT-White Detector

This article describes the event detection system of the NEXT-White detector, a 5 kg high pressure xenon TPC with electroluminescent amplification, located in the Laboratorio Subterráneo de Canfranc (LSC), Spain. The detector is based on a plane of photomultipliers (PMTs) for energy measurements and a silicon photomultiplier (SiPM) tracking plane for offline topological event filtering. The event detection system, based on the SRS-ATCA data acquisition system developed in the framework of the CERN RD51 collaboration, has been designed to detect multiple events based on online PMT signal energy measurements and a coincidence-detection algorithm. Implemented on FPGA, the system has been successfully running and evolving during NEXT-White operation. The event detection system brings some relevant and new functionalities in the field. A distributed double event processor has been implemented to detect simultaneously two different types of events thus allowing simultaneous calibration and physics runs. This special feature provides constant monitoring of the detector conditions, being especially relevant to the lifetime and geometrical map computations which are needed to correct high-energy physics events. Other features, like primary scintillation event rejection, or a double buffer associated with the type of event being searched, help reduce the unnecessary data throughput thus minimizing dead time and improving trigger efficiency.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Accelerating data acquisition with FPGA-based edge machine learning: a case study with LCLS-II

New scientific experiments and instruments generate vast amounts of data that need to be transferred for storage or further processing, often overwhelming traditional systems. Edge machine learning (EdgeML) addresses this challenge by integrating machine learning (ML) algorithms with edge computing, enabling real-time data processing directly at the point of data generation. EdgeML is particularly beneficial for environments where immediate decisions are required, or where bandwidth and storage are limited. In this paper, we demonstrate a high-speed configurable ML model in a fully customizable EdgeML system using a field programmable gate array (FPGA). Our demonstration focuses on an angular array of electron spectrometers, referred to as the ‘CookieBox,’ developed for the Linac Coherent Light Source II project. The EdgeML system captures 51.2 Gbps from a 6.4 GS s −1 analog to digital converter and is designed to integrate data pre-processing and ML inside an FPGA. Our implementation achieves an inference latency of 0.2 µs for the ML model, and a total latency of 0.4 µs for the complete EdgeML system, which includes pre-processing, data transmission, digitization, and ML inference. The modular design of the system allows it to be adapted for other instrumentation applications requiring low-latency data processing.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

A Real-Time ANPC Inverter Digital Twin with Integrated Design-For-Trust

The demand for renewable energy has increased over the last few years, and so has the demand for greater expectations within the energy market. This increasing trend has been accompanied by more significant usage of internet-connected devices (IoT), leading to critical electrical infrastructure being connected to the internet. Implementing internet connectivity with such devices and systems provides benefits such as improving the system's performance, facilitating irregularity and anomaly mitigation, and providing additional situational awareness for enhanced decision-making. However, enhancing the connected system with IoT introduces a drawback – a greater vulnerability to cyber-attacks. Cyber-attacks targeting critical infrastructure in the electrical sector have occurred in the United States and Ukraine. These cyber-attacks highlight and expose vulnerabilities that a system inherits when connecting to the internet. These attacks left thousands of customers without electricity for hours until operators could regain control of the electric utility grid. Therefore, to address the vulnerabilities of an internet-connected power electronic device, this work focused on the hardware layer of the system. Implementing a cyber-control system inside the hardware layer can significantly reduce the possibility of an attacker patching malicious controller firmware into a photovoltaic grid-connected inverter, thus mitigating the likelihood that the inverter becomes inactive a cyber-attack scenario. With this mitigation technique, if a cyberattack is successful and an attacker gains control of the network, a cyber-defense technique is in place to mitigate the impact of the cyber-attack. This additional protection layer was developed based on an innovative concept known as Digital Twin (DT). A DT, in this case, replicates an Active-Neutral Point Clamped (ANPC) inverter and was designed using a hardware language known as VHDL (Very High-Speed Integrated Circuit Hardware Description Language) and applied to Field-Programmable-GateArray (FPGA). The DT is embedded within the FPGA and contained in a controller board, the UCB (Unified Controller Board), developed by the University of Arkansas electrical engineering team. This UCB also contains two Digital Signal Processors (DSPs) responsible for generating associated signals to control an authentic physical inverter. These DSP signals are received and processed by the FPGA that implements the DT of an ANPC; in other words, it simulates in realtime the expected output of an actual ANPC inverter using the signals from the DSP. When a new firmware is ready to be patched, the DT provides output signals simulating behavior that a real ANPC inverter would generate with the new firmware. The new firmware is tested to check if it meets all the operational requirements established using a Design-For-Trust technique (DFTr). If the new firmware fails in at least one of the DFT tests, it is considered malicious and must be rejected. This work is divided into sections, such as Background, which explains the pieces that were used and the strategy behind this work; Process and Procedure, which explains the methodology that was adopted to prove the reliability and effectiveness of this work; Results and Discussion, where the simulations and results are described and explained; followed by Conclusion and Future work section, which concludes this work and adds possible future projects to continue this work further.

do Amaral Custodio, Paulo Vitor↗
DOE OSTI2024

DIRECT RF SAMPLING BASED LLRF CONTROL SYSTEM FOR C-BAND LINEAR ACCELERATOR

Low Level RF (LLRF) control systems of linear accel- erators (LINACs) are typically implemented with hetero- dyne based architectures, which have complex analog RF mixers for up and down conversion. The Gen 3 Radio Fre- quency System-on-Chip (RFSoC) device from AMD Xilinx integrates data converters with maximum RF frequency of 6 GHz. This enables direct RF sampling of C-band LLRF signal typically operated at 5.712 GHz without any analogue mixers, which can significantly simplify the system architec- ture. The data converters sample RF signals in higher order Nyquist zones and then up or down convert digitally by the integrated data path in RFSoC. The closed-loop feedback control firmware implemented in FPGA integrated in RF- SoC can process the base-band signal from the ADC data path and calculate the updated phase and amplitude to be up- mixed by the DAC data path. We have developed a C-band LLRF control RFSoC platform with direct RF sampling, which targets Cool Copper Collider (𝐶3) and other C or S band LINAC research and development projects. In this paper, the architecture of the platform will be described. We have optimized the configuration of the data converter and characterized performance of them with RF pulses. The test results for some of the key performance parameters for the LLRF platform with our custom solid-state amplifier, such as phase and amplitude stability, will be discussed in this paper.

Liu, C↗
DOE OSTI2023

Exploration of Optimizing FPGA-based Qubit Controller for Experiments on Superconducting Quantum Computing Hardware

This work explores avenues and target areas for optimizing FPGA-based control hardware for experiments conducted on superconducting quantum computing systems and serves as an introduction to some of the current research at the intersection of classical and quantum computing hardware. With the promise of building larger-scale error-corrected quantum computers based on superconducting qubit architecture, innovations to room-temperature control electronics are needed to bring these quantum realizations to fruition. The QICK (Quantum Instrumentation Control Kit) is one leading experimental FPGA-based implementations. However, its integration into other experimental quantum computing architectures, especially those using superconducting radiofrequency (SRF) cavities, is largely unexplored. We identify some key target areas for optimizing control electronics for superconducting qubit architectures and provide some preliminary results to the resolution of a control pulse waveform. With optimizations targeted at 3D superconducting qubit setups, we hope to bring to light some of the requirements in classical computational methodologies to bring out the full potential of this quantum computing architecture, and to convey the excitement of progress in this research.

43 PARTICLE ACCELERATORS↗
DOE OSTI2021

hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices

Accessible machine learning algorithms, software, and diagnostic tools for energy-efficient devices and systems are extremely valuable across a broad range of application domains. In scientific domains, real-time near-sensor processing can drastically improve experimental design and accelerate scientific discoveries. To support domain scientists, we have developed hls4ml, an open-source software-hardware codesign workflow to interpret and translate machine learning algorithms for implementation with both FPGA and ASIC technologies. We expand on previous hls4ml work by extending capabilities and techniques towards low-power implementations and increased usability: new Python APIs, quantization-aware pruning, end-to-end FPGA workflows, long pipeline kernels for low power, and new device backends include an ASIC workflow. Taken together, these and continued efforts in hls4ml will arm a new generation of domain scientists with accessible, efficient, and powerful tools for machine-learning-accelerated discovery.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

FPGA Acceleration of GCN in Light of the Symmetry of Graph Adjacency Matrix

Graph Convolutional Neural Networks (GCNs) are widely used to process large-scale graph data. Different from deep neural networks (DNNs), GCNs are sparse, irregular, and unstructured, posing unique challenges to hardware acceleration with regular processing elements (PEs). In particular, the adjacency matrix of a GCN is extremely sparse, leading to frequent but irregular memory access, low spatial/temporal data locality and poor data reuse. Furthermore, a realistic graph usually consists of unstructured data (e.g., unbalanced distributions), creating significantly different processing times and imbalanced workload for each node in GCN acceleration. To overcome these challenges, we propose an end-to-end hardware-software co-design to accelerate GCNs on resource-constrained FPGAs with the features including: (1) A custom dataflow that leverages symmetry along the diagonal of the adjacency matrix to accelerate feature aggregation for undirected graphs. We utilize either the upper or the lower triangular matrix of the adjacency matrix to perform aggregation in GCN to improve data reuse. (2) Unified compute cores for both aggregation and transform phases, with full support to the symmetry-based dataflow. These cores can be dynamically reconfigured to the systolic mode for transformation or as individual accumulators for aggregation in GCN processing. (3) Preprocessing of the graph in software to rearrange the edges and features to match the custom dataflow. This step improves the regularity in memory access and data reuse in the aggregation phase. Moreover, we quantize the GCN precision from FP32 to INT8 to reduce the memory footprint without losing the inference accuracy. We implement our accelerator design in Intel Stratix10 MX FPGA board with HBM2, and demonstrate 1.3x-110.5x improvement in end-to-end GCN latency as compared to the state-of the-art FPGA implementations, on the graph datasets of Cora, Pubmed, Citeseer and Reddit.

Nair, Gopikrishnan R.↗
DOE OSTI2024

End-to-end codesign of Hessian-aware quantized neural networks for FPGAs

Here, we develop an end-to-end workflow for the training and implementation of co-designed neural networks (NNs) for efficient field-programmable gate array (FPGA) hardware. Our approach leverages Hessian-aware quantization of NNs, the Quantized Open Neural Network Exchange intermediate representation, and the hls4ml tool flow for transpiling NNs into FPGA firmware. This makes efficient NN implementations in hardware accessible to nonexperts in a single open sourced workflow that can be deployed for real-time machine-learning applications in a wide range of scientific and industrial settings. We demonstrate the workflow in a particle physics application involving trigger decisions that must operate at the 40-MHz collision rate of the CERN Large Hadron Collider (LHC). Given the high collision rate, all data processing must be implemented on FPGA hardware within the strict area and latency requirements. Based on these constraints, we implement an optimized mixed-precision NN classifier for high-momentum particle jets in simulated LHC proton-proton collisions.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

Proton pulse charge calculation algorithm in Beam Power Limiting System at Spallation Neutron Source

A proton pulse charge calculation algorithm in the Beam Power Limiting System (BPLS) at the Spallation Neutron Source (SNS) was developed and implemented in an FPGA. The algorithm calculates one-minute running average of the pulse charges and issues a fault to the Personal Protection System (PPS) and the Machine Protection System (MPS) when a limit is reached. A bit-accurate model of the algorithm was first developed and tested in Matlab® and then implemented and simulated in VHDL using Vivado® design environment. Finally, the algorithm was verified on a µTCA-based hardware platform.

Bobrek, Miljko [ORNL] (ORCID:0000000332763451)↗
DOE OSTI2026

ML–Enabled FPGA Framework for Fast Quantum State Discrimination in Mid-Circuit Measurement Regimes

Accurate and low-latency quantum state discrimination is essential for protocols involving mid-circuit measurement (MCM) and conditional feed-forward. In superconducting quantum systems, conventional readout pipelines transfer measurement data to host processors for post-processing, introducing millisecond-scale delays that far exceed qubit coherence times. To overcome this bottleneck, we present an in-situ machine learning (ML) inference engine implemented on an FPGA for real-time quantum state discrimination. Our design performs inference directly on digitized readout signals with 40 ns latency, supports both qubit and qutrit readout, and enables conditional operations without host-side intervention. This capability is critical for MCM and for feedback-driven protocols such as quantum error correction. We validate the system on superconducting transmon hardware, demonstrating robust discrimination fidelity across multiple qubit and qutrit channels. We further demonstrate conditional qutrit logic driven by FPGA-resident classification, highlighting the potential of low-latency ML-on-FPGA control for NISQ applications and scalable fault-tolerant quantum computing.

Vora, Neel [Lawrence Berkeley National Laboratory ↗