Remote attackers can recover "FPGA pentimento" - long-removed data belonging to a prior user or proprietary design image on a cloud FPGA. Just as a pentimento of a painting can be exposed by infrared imaging, FPGA pentimentos can be exposed by signal timing sensors. The data constituting an FPGA pentimento is imprinted on the device through bias temperature instability effects on the underlying transistors. Measuring this degradation using a time-to-digital converter allows an attacker to (1) extract proprietary details or keys from an encrypted FPGA design image available on the AWS marketplace and (2) recover information from a previous user of a cloud-FPGA. These threat models are validated on AWS F1, with successful AES key recovery under one model.

Design space exploration (DSE) provides intelligent methods to tune the large number of optimization parameters present in modern FPGA high-level synthesis tools. High-level synthesis parameter tuning is a time-consuming process due to lengthy hardware compilation times—synthesizing an FPGA design can take tens of hours. DSE helps find an optimal solution faster than brute-force methods without relying on designer intuition to achieve high-quality results. Sherlock is a DSE framework that can handle multiple conflicting optimization objectives and aggressively focuses on finding Pareto-optimal solutions. Sherlock integrates a model selection process to choose the regression model that helps reach the optimal solution faster. Sherlock designs a strategy based around the multi-armed bandit problem, opting to balance exploration and exploitation based on the learned and expected results. Sherlock can decrease the importance of models that do not provide correct estimates, reaching the optimal design faster. Sherlock is capable of tailoring its choice of regression models to the problem at hand, leading to a model that best reflects the application design space. We have tested the framework on a large dataset of FPGA design problems and found that Sherlock converges toward the set of optimal designs faster than similar frameworks.

Domain-specific accelerators obtain performance benefits by restricting their algorithmic domain. These accelerators utilize specialized languages constrained to particular hardware, thus trading off expressiveness for high performance. The pendulum has swung from one hardware for all domains (general-purpose processors) to one hardware per individual domain. The middle-ground on this spectrum-which provides a unified computational stack across multiple, but not all, domains- is an emerging and open research challenge. This paper sets out to explore this region and its associated tradeoff between expressiveness and performance by defining a cross-domain stack, dubbed PolyMath. This stack defines a high-level cross-domain language (CDL), called PMLang, that in a modular and reusable manner encapsulates mathematical properties to be expressive across multiple domains-Robotics, Graph Analytics, Digital Signal Processing, Deep Learning, and Data Analytics. PMLang is backed by a recursively-defined intermediate representation allowing simultaneous access to all levels of operation granularity, called sr DFG. Accelerator-specific or domain-specific IRs commonly capture operations in the granularity that best fits a set of Domain-Specific Architectures (DSAs). In contrast, the recursive nature of the sr DFG enables simultaneous access to all the granularities of computation for every operation, thus forming an ideal bridge for converting to various DSA-specific IRs across multiple domains. Our stack unlocks multi-acceleration for end-to-end applications that cross the boundary of multiple domains each comprising different data and compute patterns. Evaluations show that by using PolyMath it is possible to harness accelerators across the five domains to realize an average speedup of 3.3× over a Xeon CPU along with 18.1× reduction in energy. In comparison to Jetson Xavier and Titan XP, cross-domain acceleration offers 1.7× and 7.2× improvement in performance-per-watt, respectively. We measure the cross-domain expressiveness and performance tradeoff by comparing each benchmark against its hand-optimized implementation to achieve 83.9% and 76.8% of the optimal performance for single-domain algorithms and end-to-end applications. For the two case studies of end-to-end applications (comprising algorithms from multiple domains), results show that accelerating all kernels offers an additional 2.0× speedup over CPU, 6.1× improvement in performance-per-watt over Titan Xp, and 2.8× speedup over Jetson Xavier compared to only the one most effective single-domain kernel being accelerated. Finally, we examine the utility and expressiveness of PolyMath through a user study, which shows, on average, PolyMath requires 1.9× less time to implement algorithms from two different domains with 2.5× fewer lines of code relative to Python.

An adversary can exploit side-channel information such as power consumption, electromagnetic (EM) emanations, acoustic vibrations or the timing of encryption operations to derive the secret key from an electronic device. Signature aTtenuation Embedded CRYPTO with Low-Level metAl Routing (STELLAR) is a technique to mitigate power and EM-based attacks, however, it incurs 50% power overhead. This work presents iSTELLAR, which reduces the power overhead by operating STELLAR intermittently utilizing an intelligent scheduling algorithm. The proposed scheduling algorithm for iSTELLAR determines the optimal locations during the crypto operation to turn STELLAR ON, and thereby reduces the power overhead by$> 30\%$compared to the normal STELLAR operation, while eliminating the information leakage.

Although hardware security has received significant attention in the past decade or so, security design and validation engineers and researchers in industry, academia, and government have not still been equipped with a mature security-aware toolset to automatically and effectively analyze designs for various types of security vulnerabilities at different to detect and fix the security issues or build security in designs efficiently and easily. Despite such a demand, currently, there is not an ecosystem of security-aware Electronic Design Automation (EDA) or Computer-Aided Design (CAD) tools whereas the commercial design for security and validation tools are still in their infancy. However, there exist many research works that try to come up with security analysis engines and provide solutions to address different classes of security issues such as data leakage, access control violation, side-channel leakage, hardware Trojans and malicious changes, and vulnerabilities to physical attacks, fault-injection attacks, reverse engineering attacks, and chip counterfeiting or overproduction attacks. This paper presents the foundation established by several academic and industry researchers who have been supporting the realization of an ecosystem of security-aware CAD tools with their focus on hardware security coverage and fault-injection assessment for SoC designs, and security assurance standardization for electronic design integration.

Simultaneous Localization And Mapping (SLAM) is an important technique used in robotics, computer vision, and virtual/augmented reality. SLAM algorithms have moved past creating sparse maps to making dense 3D reconstruction of the environment. Dense SLAM algorithms have high computational demands that require hardware acceleration to be done efficiently in real-time. FPGAs are an attractive compute platform for SLAM systems as they are low power and high performance. Unfortunately, dense SLAM algorithms are complex and FPGAs are notoriously difficult to program. In this work, we study the best techniques for accelerating 3D reconstruction on FPGA. We analyze a 3D reconstruction system, and implement modular FPGA designs for the main components of this application. We target both an FPGA SoC and a larger FPGA PCIe board, and perform a design space exploration (DSE) of our designs. We analyze the results of our DSE, characterize the design spaces to highlight important features, and we implement the best designs in an open-source and end-to-end dense SLAM system running on a FPGA SoC board. On the SoC board, using the FPGA increases the throughput of the whole application by a factor of two compared to the ARM processor, and individual algorithms are up to 38 times faster on the FPGA.

For many devices, power side-channel attacks are an effective means of obtaining secret keys from cryptographic algorithms. Recently, methods have been proposed to assess the vulnerability of devices to these attacks. While existing approaches effectively evaluate device vulnerability to attacks at specific points during execution, they do not consider the power measurement vectors holistically, using all time points in the measurement. This is necessary in order to accurately assess resistance to multi-target attacks. In this work, we identify characteristics of an ideal holistic side-channel security metric and develop a metric under these criteria. We demonstrate that our approach correctly ranks different FPGA implementations of AES with respect to attack difficulty.

As demonstrated by numerous practical attacks, the physical act of computation emits unintended and damaging information through infinitesimal variations in timing, power, and resource contention. While there are many techniques for preventing the leakage of information through power channels for specific cryptographic units, they are typically either built directly into the hardware logic or exploit intricate mathematical properties of the algorithm itself. However, such leaks are not uniform in time but, as we show, rather occur in specific bursts. Exploiting this observation we propose a set of software-controlled techniques allowing for the seamless disconnection and reconnection of general purpose programmable components in a system-on-chip. Such a system is capable of providing brief moments of electrical isolation during which the most critical computations can be performed free from both timing and power measurement. Of course, disconnection comes at a cost. To balance the resulting trade-off between overhead and security effectively, we describe a new analysis technique to uncover the "leakiest" intervals of time, we provide an algorithm to co-optimize the covering of these intervals and the performance/energy costs under a set of architecture imposed constraints, and explore the architectural and software ramifications of such intermittent disconnection. In the end we find that by hiding only between 15% and 30% of the trace, at a performance cost of between 15% and 50%, we are able to reduce the mutual information between the leakage model and key bits by 75% on average, and to nearly zero in specific cases.

Cryptographic cores are known to leak information about their private key due to runtime variations, and there are many well-known attacks that can exploit this timing channel. In this paper, we study how information theoretic measures can quantify the amount of key leakage that can be exacted from runtime measurements. We develop and analyze 22 Rivest-Shamir-Adleman (RSA) hardware designs-each with unique performance optimizations, timing channel mitigation techniques, or discretization/randomization countermeasures. We demonstrate the effectiveness of information theoretic measures for quantifying timing leakage through correlation analysis of information theoretic measurements and attack results. Experimental results show that mutual information is a promising technique for quantifying timing leakage for RSA, advanced encryption standard, and elliptic curve cryptography ciphers, i.e., the mutual information correlates to being able to successfully guess the value of the private key. This is an important step toward a hardware security metric which allows designers to reason about security alongside traditional hardware design metrics like area, performance, and power.

State-of-the-art C/C++ synthesis tools lack abstractions and conveniences that are pervasive in modern software languages. Higher-order functions are particularly important as they increase productivity by concisely representing common design patterns. Providing these in hardware design environments would improve the accessibility of hardware tools for software engineers by providing familiar interfaces and abstractions. We have created an open-source library of higher-order functions synthesizable in C/C++ hardware development tools. We implement six common algorithms on a PYNQ board and conclude that our library produces results that are generally statistically indistinguishable from nonrecursive techniques.

Learning cumulative distribution functions (CDFs) is a widely studied problem in data stream summarization. While current techniques have efficient software implementations, their efficiency depends on updates to data structures that are not easily adapted to FPGA or ASIC implementation. In this work, we develop an algorithm and a compact hardware architecture for learning the CDF of a data stream and apply our technique to the problem of on-chip run-time testing for bias in the output of random number generators (RNGs). Unlike previous approaches, our method is successful regardless of the expected output distribution of the RNG under test.

Data clustering is a fundamental challenge in data analytics. It is the main task in exploratory data mining and a core technique in machine learning. As the volume, variety, velocity, and variability of data grows, we need more efficient data analysis methods that can scale towards increasingly large and high dimensional data sets. We develop a streaming clustering algorithm that is highly amenable to hardware acceleration. Our algorithm eliminates the need to store the data objects, which removes limits on the size of the data that we can analyze. Our algorithm is highly parameterizable, which allows it to fit to the characteristics of the data set, and scale towards the available hardware resources. Our streaming hardware core can handle more than 40 Msamples/s when processing 3-dimensional streaming data and up to 1.78 Msamples/s for 70-dimensional data. To validate the accuracy and performance of our algorithms we compare it with several common clustering techniques on several different applications. The experimental result shows that it outperforms other prior hardware accelerated clustering systems.

High-level synthesis tools allow programmers to use OpenCL to create FPGA designs. Unfortunately, these tools have a complex compilation process that can take several hours to synthesize a single design. This creates a significant barrier for design optimization since even experts typically need to test many designs due to the non-obvious interactions between the different optimizations. Thus, understanding the design space, and guiding the optimization process is a crucial requirement for enabling the widespread adoption of these high-level synthesis tools. However this requires a significant amount of design space data that is currently unavailable or difficult to generate. To solve this problem, we present an OpenCL FPGA benchmark suite. We outfitted each benchmark with a range of optimization parameters (or knobs), compiled over 8300 unique designs using the Altera OpenCL SDK, executed them on a Terasic DE5 board, and recorded their corresponding performance and utilization characteristics. We describe the resulting design spaces, and perform a statistical analysis of the optimization configurations which provides valuable architecture insights to FPGA developers. We make the benchmarks and results completely open-source to give opportunities for the community to perform additional analyses and provide a repository of well-documented designs for follow-on research.

Nonnegative least squares (NNLS) optimization is an important algorithmic component of many problems in science and engineering, including image segmentation, spectral deconvolution, and reconstruction of compressively sensed data. Each of these areas can benefit from high performance implementations suitable for embedded applications. Unfortunately, the NNLS problem has no solution expressible in closed-form, and many popular algorithms are not amenable to compact and scalable hardware implementation. Classical iterative algorithms generally have a per-iteration computational cost with cubic growth, and interdependencies which limit parallel approaches. In this paper we develop two efficient hardware architectures. One is based on a novel algorithm we develop in this paper specifically to reduce FPGA area consumption while preserving performance. We implement our architectures on a very small FPGA and apply them to reconstruction of a compressively sensed signal showing residual error results competitive with traditional algorithms.

Cryptographic function implementations are known to leak information about private keys through timing information. By using statistical analysis of the variations in runtime required to encrypt different messages, an attacker can relatively easily determine the key with high probability. There are many mitigation techniques to combat these side channels; however, there are limited metrics available to quantify the effectiveness of these mitigation attacks. In this work, we employ information theoretic ideas to quantify the amount of leakage that can be extracted from runtime measurements and reveal the influence of individual key bits on the timing observations across a variety of hardware implementations. By studying different RSA hardware architectures (each with different performance optimizations and mitigation techniques), we determine the effectiveness of these information theoretic techniques against the success of attacks. Our experimental results show that mutual information is a promising metric to quantify timing-based information leakage and it also correlates to the attack-ability of a cryptographic implementation.