Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Field Programmable Gate Arrays”

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 289 records · Page 16

FPGA Trigger System to Run Klystrons

The Klystron Department is in need of a new trigger system to update the laboratory capabilities. The objective of the research is to develop the trigger system using Field Programmable Gate Array (FPGA) technology with a user interface that will allow one to communicate with the FPGA via a Universal Serial Bus (USB). This trigger system will be used for the testing of klystrons. The key materials used consists of the Xilinx Integrated Software Environment (ISE) Foundation, a Programmable Read Only Memory (Prom) XCF04S, a Xilinx Spartan 3E 35S500E FPGA, Xilinx Platform Cable USB II, a Printed Circuit Board (PCB), a 100 MHz oscillator, and an oscilloscope. Key considerations include eight triggers, two of which have variable phase shifting capabilities. Once the project was completed the output signals were able to be manipulated via a Graphical User Interface by varying the delay and width of the signal. This was as planned; however, the ability to vary the phase was not completed. Future work could consist of being able to vary the phase. This project will give the operators in the Klystron Department more flexibility to run various tests.

43 PARTICLE ACCELERATORS↗

Graph Neural Networks for Charged Particle Tracking on FPGAs

The determination of charged particle trajectories in collisions at the CERN Large Hadron Collider (LHC) is an important but challenging problem, especially in the high interaction density conditions expected during the future high-luminosity phase of the LHC (HL-LHC). Graph neural networks (GNNs) are a type of geometric deep learning algorithm that has successfully been applied to this task by embedding tracker data as a graph—nodes represent hits, while edges represent possible track segments—and classifying the edges as true or fake track segments. However, their study in hardware- or software-based trigger applications has been limited due to their large computational cost. In this paper, we introduce an automated translation workflow, integrated into a broader tool called hls4ml , for converting GNNs into firmware for field-programmable gate arrays (FPGAs). We use this translation tool to implement GNNs for charged particle tracking, trained using the TrackML challenge dataset, on FPGAs with designs targeting different graph sizes, task complexites, and latency/throughput requirements. This work could enable the inclusion of charged particle tracking GNNs at the trigger level for HL-LHC experiments.

Elabd, Abdelrahman↗

Entropy Analysis of FPGA Interconnect and Switch Matrices for Physical Unclonable Functions

Random variations in microelectronic circuit structures represent the source of entropy for physical unclonable functions (PUFs). In this paper, we investigate delay variations that occur through the routing network and switch matrices of a field-programmable gate array (FPGA). The delay variations are isolated from other components of the programmable logic, e.g., look-up tables (LUTs), flip-flops (FFs), etc., using a feature of Xilinx FPGAs called dynamic partial reconfiguration (DPR). A set of partial designs is created to fix the placement of a time-to-digital converter (TDC) and supporting infrastructure to enable the path delays through the target interconnect and switch matrices to be extracted by subtracting out common-mode delay components. Delay variations are analyzed in the different levels of routing resources available within FPGAs, i.e., local routing and across-chip routing. Data are collected from a set of Xilinx Zynq 7010 devices, and a statistical analysis of within-die variations in delay through a set of the randomly-generated and hand-crafted interconnects is presented.

97 MATHEMATICS AND COMPUTING↗

Real-time Anomaly Detection for Liquid Argon Time Projection Chambers

We present a real-time anomaly detection framework for liquid argon time projection chambers (LArTPCs), targeting applications in particle physics experiments such as the Short Baseline Near Detector (SBND) or the future Deep Underground Neutrino Experiment (DUNE). These experiments employ detectors that generate and stream high-resolution but sparse images of neutrino and other particle interactions. Our approach utilizes anomaly detection with autoencoders, compressed through knowledge distillation (KD), to enable the detection of anomalous signals in the data through efficient inference on resource-constrained hardware. The framework is targeted for deployment on computing platforms equipped with field-programmable gate arrays (FPGAs), GPUs, or CPUs, allowing low-latency selection of relevant activity directly from the raw detector data stream. We demonstrate that our approach is suitable for the detection and localization of anomalously "high-multiplicity" activity, and outline promising applications for LArTPC online data filtering and triggering.

FOS: Physical sciences↗

Rapid Inference of Logic Gate Neural Networks for Anomaly Detection in High Energy Physics

The increasing data rates and complexity of detectors at the Large Hadron Collider (LHC) necessitate fast and efficient machine learning models, particularly for rapid selection of what data to store, known as triggering. Building on recent work in differentiable logic gates, we present a public implementation of a Convolutional Differentiable Logic Gate Neural Network (CLGN). We apply this to detecting anomalies at the Level-1 Trigger at CMS using public data from the CICADA project. We demonstrate that the CLGN achieves physics performance on par with or superior to conventional quantized neural networks. We also synthesize an LGN for a Field-Programmable Gate Array (FPGA) and show highly promising FPGA characteristics, notably zero Digital Signal Processor (DSP) resource usage. This work highlights the potential of logic gate networks for high-speed, on-detector inference in High Energy Physics and beyond.

FOS: Physical sciences↗

hls4ml: A Flexible, Open-Source Platform for Deep Learning Acceleration on Reconfigurable Hardware

We present hls4ml, a free and open-source platform that translates machine learning (ML) models from modern deep learning frameworks into high-level synthesis (HLS) code that can be integrated into full designs for field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs). With its flexible and modular design, hls4ml supports a large number of deep learning frameworks and can target HLS compilers from several vendors, including Vitis HLS, Intel oneAPI and Catapult HLS. Together with a wider eco-system for software-hardware co-design, hls4ml has enabled the acceleration of ML inference in a wide range of commercial and scientific applications where low latency, resource usage, and power consumption are critical. In this paper, we describe the structure and functionality of the hls4ml platform. The overarching design considerations for the generated HLS code are discussed, together with selected performance results.

FOS: Computer and information sciences↗

Cross strip microchannel plate imaging photon counters with high time resolution

We have implemented cross strip readout microchannel plate detectors in 18 mm active area format including open face (UV/particle) and sealed tube (optical) configurations. These have been tested with a field programmable gate array based parallel channel electronics for event encoding which can process high input event rates (> 5 MHz) with high spatial resolution. Using small pore MCPs (6 {micro}m) operated in a pair, we achieve gains of >5 x 10{sup 5} which is sufficient to provide spatial resolution of <35 {micro}m FHWM, with self triggered event timing accuracy of {approx}2 ns for sealed tube optical sensors. A peak quantum efficiency of {approx}19% at 500 nm has been achieved with SuperGenII photocathodes that have response over the 400 nm to 900 nm range. Local area counting rates of up to >200 events/mcp pore sec{sup -1} have been attained, along with image linearity and stability to better than 50 {micro}m.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Efficient Graph Based Assembly of Short-Read Sequences on Hybrid Core Architecture

Advanced architectures can deliver dramatically increased throughput for genomics and proteomics applications, reducing time-to-completion in some cases from days to minutes. One such architecture, hybrid-core computing, marries a traditional x86 environment with a reconfigurable coprocessor, based on field programmable gate array (FPGA) technology. In addition to higher throughput, increased performance can fundamentally improve research quality by allowing more accurate, previously impractical approaches. We will discuss the approach used by Convey?s de Bruijn graph constructor for short-read, de-novo assembly. Bioinformatics applications that have random access patterns to large memory spaces, such as graph-based algorithms, experience memory performance limitations on cache-based x86 servers. Convey?s highly parallel memory subsystem allows application-specific logic to simultaneously access 8192 individual words in memory, significantly increasing effective memory bandwidth over cache-based memory systems. Many algorithms, such as Velvet and other de Bruijn graph based, short-read, de-novo assemblers, can greatly benefit from this type of memory architecture. Furthermore, small data type operations (four nucleotides can be represented in two bits) make more efficient use of logic gates than the data types dictated by conventional programming models.JGI is comparing the performance of Convey?s graph constructor and Velvet on both synthetic and real data. We will present preliminary results on memory usage and run time metrics for various data sets with different sizes, from small microbial and fungal genomes to very large cow rumen metagenome. For genomes with references we will also present assembly quality comparisons between the two assemblers.

59 BASIC BIOLOGICAL SCIENCES↗

Fast BPM data distribution for global orbit feedback using commercial gigabit ethernet technology

In order to correct beam perturbations in RHIC around 10Hz, a new fast data distribution network was required to deliver BPM position data at rates several orders of magnitude above the capability of the existing system. The urgency of the project limited the amount of custom hardware that could be developed, which dictated the use of as much commercially available equipment as possible. The selected architecture uses a custom hardware interface to the existing RHIC BPM electronics together with commercially available Gigabit Ethernet switches to distribute position data to devices located around the collider ring. Using the minimum Ethernet packet size and a field programmable gate array (FPGA) based state machine logic instead of a software based driver, real-time and deterministic data delivery is possible using Ethernet. The method of adapting this protocol for low latency data delivery, bench testing of Ethernet hardware, and the logic to construct Ethernet packets using FPGA hardware will be discussed. A robust communications system using almost all commercial off-the-shelf equipment was developed in under a year which enabled retrofitting of the existing RHIC BPM system to provide 10 KHz data delivery for a global orbit feedback scheme using 72 BPMs. Total latencies from data acquisition at the BPMs to delivery at the controller modules, including very long transmission distances, were kept under 100 {micro}s, which provide very little phase error in correcting the 10 Hz oscillations. Leveraging off of the speed of Gigabit Ethernet and wide availability of Ethernet products enabled this solution to be fully implemented in a much shorter time and at lower cost than if a similar network was developed using a proprietary method.

43 PARTICLE ACCELERATORS↗

A high performance DAC /DDS daughter module for the RHIC LLRF platform

The RHIC LLRF upgrade is a flexible, modular system. Output signals are generated by a custom designed XMC card with 4 high speed digital to analog (DAC) converters interfaced to a high performance field programmable gate array (FPGA). This paper discusses the hardware details of the XMC DAC board as well as the implementation of a low noise rf synthesizer with digital IQ modulation. This synthesizer also provides injection phase cogging and frequency hop rebucketing capabilities. A new modular RHIC LLRF system was recently designed and commissioned based on custom designed XMC cards. As part of that effort a high speed, four channel DAC board was designed. The board uses Maxim MAX5891 16 bit DACs with a maximum update rate of 600 Msps. Since this module is intended to be used for many different systems throughout the Collider Accelerator complex, it was designed to be as generic as possible. One major application of this DAC card is to implement digital synthesizers to provide drive signals to the various cavities at RHIC. Since RHIC is a storage ring with stores that typically last many hours, extremely low RF noise is a critical requirement. Synchrotron frequencies at RHIC range from a few hertz to several hundred hertz depending on the species and point in the acceleration cycle so close in phase noise is a major concern. The RHIC LLRF system uses the Update Link, a deterministic, high speed data link that broadcasts the revolution frequency and the synchronous phase angle. The digital synthesizers use this data to generate a properly phased analog drive signal. The synthesizers must also provide smooth phase shifts for cogging and support frequency shift rebucketing. One additional feature implemented in the FPGA is a digital waveform generator (WFG) that generates I and Q data pairs based on a user selected amplitude and phase profile as a function of time.

43 PARTICLE ACCELERATORS↗

A bunch to bucket phase detector for the RHIC LLRF upgrade platform

As part of the overall development effort for the RHIC LLRF Upgrade Platform [1,2,3], a generic four channel 16 bit Analog-to-Digital Converter (ADC) daughter module was developed to provide high speed, wide dynamic range digitizing and processing of signals from DC to several hundred megahertz. The first operational use of this card was to implement the bunch to bucket phase detector for the RHIC LLRF beam control feedback loops. This paper will describe the design and performance features of this daughter module as a bunch to bucket phase detector, and also provide an overview of its place within the overall LLRF platform architecture as a high performance digitizer and signal processing module suitable to a variety of applications. In modern digital control and signal processing systems, ADCs provide the interface between the analog and digital signal domains. Once digitized, signals are then typically processed using algorithms implemented in field programmable gate array (FPGA) logic, general purpose processors (GPPs), digital signal processors (DSPs) or a combination of these. For the recently developed and commissioned RHIC LLRF Upgrade Platform, we've developed a four channel ADC daughter module based on the Linear Technology LTC2209 16 bit, 160 MSPS ADC and the Xilinx V5FX70T FPGA. The module is designed to be relatively generic in application, and with minimal analog filtering on board, is capable of processing signals from DC to 500 MHz or more. The module's first application was to implement the bunch to bucket phase detector (BTB-PD) for the RHIC LLRF system. The same module also provides DC digitizing of analog processed BPM signals used by the LLRF system for radial feedback.

43 PARTICLE ACCELERATORS↗

New Beam Loss Monitor for 12 GeV Upgrade

This paper describes a new VME based machine protection Beam Loss Monitor (BLM) signal processing board designed at Jefferson Lab to replace the current CAMAC based BLM board. The new eight-channel BLM signal processor has linear, logarithmic, and integrating amplifiers that simultaneously provide the optimal signal processing for each application. Amplified signals are digitized and then further processed through a Field Programmable Gate Array (FPGA). Combining both the diagnostic and machine protection functions in each channel allows the operator to tune-up and monitor beam operations while the machine protection is integrating the same signal. Other features include extensive built-in-self-test, fast shutdown interface (FSD), and 16-Mbit buffers for beam loss transient play-back. The new VME BLM board features high sensitivity, high resolution, and low cost per channel.

43 PARTICLE ACCELERATORS↗

Commissioning of the IGp Feedback System at DAFNE

The iGp (Integrated Gigasample Processor) is an innovative digital bunch-by-bunch feedback system developed by a KEK / SLAC / INFN-LNF joint collaboration. The processing unit can sample at 500 MHz and compute the bunch-by-bunch output signal for up to {approx}5000 bunches. The feedback gateware code is implemented inside just one FPGA (Field Programmable Gate Array) chip, a Xilinx Virtex-II. The FPGA implements two banks of 16-tap FIR (Finite Impulse Response) filters. Each filter is realtime programmable through the operator interface. At DA{Phi}NE, the Frascati {Phi}-Factory, two iGp units have been commissioned in the April 2007. The iGp systems have substituted the previous betatron feedback systems. This insertion has been very fast and has shown no problems involving just a substitution of the old, less flexible, digital systems, letting unchanged the baseband analog frontend and backend. The commissioning has been very simple, due to the complete and powerful EPICS operator interface, working well in local and remote operations. The software includes also tools for analyzing post processor data. A description of the commissioning with the operations done is reported.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

A low-power wave union TDC implemented in FPGA

A low-power time-to-digital convertor (TDC) for an application inside a vacuum has been implemented based on the Wave Union TDC scheme in a low-cost field programmable gate array (FPGA) device. Bench top tests have shown that a time measurement resolution better than 30 ps (standard deviation of time differences between two channels) is achieved. Special firmware design practices are taken to reduce power consumption. The measurements indicate that with 32 channels fitting in the FPGA device, the power consumption on the FPGA core voltage is approximately 9.3 mW/channel and the total power consumption including both core and I/O banks is less than 27 mW/channel.

43 PARTICLE ACCELERATORS↗

SPALLATION NEUTRON SOURCE HIGH-POWER PROTECTION MODULE TEST STAND

The Spallation Neutron Source (SNS) High-Power Protection Module (HPM) provides interlocks and fast shutdown for the radio frequency (RF) system to protect the accelerating structures and high power RF (HPRF) Distribution System. The HPM has required some functional upgrades since the start of beam operations and an upgrade to the HPM test stand was required to support these added features. The HPM test stand currently verifies functionality, RF channel calibration, and measurement of the speed of shutdown to ensure the specifications are met. The upgraded test stand was implemented in a Field Programmable Gate Array (FPGA) to allow for future growth and flexibility. Work is currently progressing on automation of the test stand to better perform the required module calibration schedule.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Advanced High-Speed Framing Camera Development for Fast, Visible Imaging Experiments

The advances in high-voltage switching developed in this project allow a camera user to rapidly vary the number of output frames from 1 to 25. A high-voltage, variable-amplitude pulse train shifts the deflection location to the new frame location during the interlude between frames, making multiple frame counts and locations possible. The final deflection circuit deflects to five different frame positions per axis, including the center position, making for a total of 25 frames. To create the preset voltages, electronically adjustable {+-}500 V power supplies were chosen. Digital-to-analog converters provide digital control of the supplies. The power supplies are clamped to {+-}400 V so as not to exceed the voltage ratings of the transistors. A field-programmable gated array (FPGA) receives the trigger signal and calculates the combination of plate voltages for each frame. The interframe time and number of frames are specified by the user, but are limited by the camera electronics. The variable-frame circuit shifts the plate voltages of the first frame to those of the second frame during the user-specified interframe time. Designed around an electrostatic image tube, a framing camera images the light present during each frame (at the photocathode) onto the tube’s phosphor. The phosphor persistence allows the camera to display multiple frames on the phosphor at one time. During this persistence, a CCD camera is triggered and the analog image is collected digitally. The tube functions by converting photons to electrons at the negatively charged photocathode. The electrons move quickly toward the more positive charge of the phosphor. Two sets of deflection plates skew the electron’s path in horizontal and vertical (x axis and y axis, respectively) directions. Hence, each frame’s electrons bombard the phosphor surface at a controlled location defined by the voltages on the deflection plates. To prevent the phosphor from being exposed between frames, the image tube is gated off between exposures.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Advanced High-Speed 16-Bit Digitizer System

The fastest commercially available 16-bit ADC can only perform around 200 mega-samples per second (200 MS/s). Connecting ADC chips together in eight different time domains increases the quantity of samples taken by a factor of eight. This method of interleaving requires that the input signal being sampled is split into eight identical signals and arrives at each ADC chip at the same point in time. The splitting of the input signal is performed in the analog front end containing a wideband filter that impedance matches the input signal to the ADC chips. Each ADC uses a clock to tell it when to perform a conversion. Using eight unique clocks spaced in 45-degree increments is the method used to time shift when each ADC chip performs its conversion. Given that this control clock is a fixed frequency, the clock phase shifting is accomplished by tightly controlling the distance that the clock must travel, resulting in a time delay. The interleaved ADC chips will now generate digital data in eight different time domains. These data are processed inside a field-programmable gate array (FPGA) to move the data back into a single time domain and store it into memory. The FPGA also contains a Nios II processor that provides system control and data retrieval via Ethernet.

42 ENGINEERING↗

Efficient Aho-Corasick String Matching on Emerging Multicore Architectures

String matching algorithms are critical to several scientific fields. Beside text processing and databases, emerging applications such as DNA protein sequence analysis, data mining, information security software, antivirus, ma- chine learning, all exploit string matching algorithms [3]. All these applica- tions usually process large quantity of textual data, require high performance and/or predictable execution times. Among all the string matching algorithms, one of the most studied, especially for text processing and security applica- tions, is the Aho-Corasick algorithm. 1 2 Book title goes here Aho-Corasick is an exact, multi-pattern string matching algorithm which performs the search in a time linearly proportional to the length of the input text independently from pattern set size. However, depending on the imple- mentation, when the number of patterns increase, the memory occupation may raise drastically. In turn, this can lead to significant variability in the performance, due to the memory access times and the caching effects. This is a significant concern for many mission critical applications and modern high performance architectures. For example, security applications such as Network Intrusion Detection Systems (NIDS), must be able to scan network traffic against very large dictionaries in real time. Modern Ethernet links reach up to 10 Gbps, and malicious threats are already well over 1 million, and expo- nentially growing [28]. When performing the search, a NIDS should not slow down the network, or let network packets pass unchecked. Nevertheless, on the current state-of-the-art cache based processors, there may be a large per- formance variability when dealing with big dictionaries and inputs that have different frequencies of matching patterns. In particular, when few patterns are matched and they are all in the cache, the procedure is fast. Instead, when they are not in the cache, often because many patterns are matched and the caches are continuously thrashed, they should be retrieved from the system memory and the procedure is slowed down by the increased latency. Efficient implementations of string matching algorithms have been the fo- cus of several works, targeting Field Programmable Gate Arrays [4, 25, 15, 5], highly multi-threaded solutions like the Cray XMT [34], multicore proces- sors [19] or heterogeneous processors like the Cell Broadband Engine [35, 22]. Recently, several researchers have also started to investigate the use Graphic Processing Units (GPUs) for string matching algorithms in security applica- tions [20, 10, 32, 33]. Most of these approaches mainly focus on reaching high peak performance, or try to optimize the memory occupation, rather than looking at performance stability. However, hardware solutions supports only small dictionary sizes due to lack of memory and are difficult to customize, while platforms such as the Cell/B.E. are very complex to program.

Tumeo, Antonino↗