Using high-level synthesis techniques, this paper proposes an adaptable high-performance streaming dataflow engine for sparse matrix dense vector multiplication (SpMV) suitable for embedded FPGAs. As the SpMV is a memory-bound algorithm, this engine combines the three concepts of loop pipelining, dataflow graph, and data streaming to utilize most of the memory bandwidth available to the FPGA. The main goal of this paper is to show that FPGAs can provide comparable performance for memory-bound applications to that of the corresponding CPUs and GPUs but with significantly less energy consumption. The experimental results indicate that the FPGA provides higher performance compared to that of embedded GPUs for small and medium-size matrices by an average factor of 3.25 whereas the embedded GPU is faster for larger size matrices by an average factor of 1.58. In addition, the FPGA implementation is more energy efficient for the range of considered matrices by an average factor of 8.9 compared to the embedded CPU and GPU. A case study based on adapting the proposed SpMV optimization to accelerate the support vector machine (SVM) algorithm, one of the successful classification techniques in the machine learning literature, justifies the benefits of utilizing the proposed FPGA-based SpMV compared to that of the embedded CPU and GPU. The experimental results show that the FPGA is faster by an average factor of 1.7 and consumes less energy by an average factor of 6.8 compared to the GPU.

Heterogeneous chips that combine CPUs and FPGAs can distribute processing so that the algorithm tasks are mapped onto the most suitable processing element. New software-defined high-level design environments for these chips use general purpose languages such as C++ and OpenCL for hardware and interface generation without the need for register transfer language expertise. These advances in hardware compilers have resulted in significant increases in FPGA design productivity. In this paper, we investigate how to enhance an existing software-defined framework to reduce overheads and enable the utilization of all the available CPU cores in parallel with the FPGA hardware accelerators. Instead of selecting the best processing element for a task and simply offloading onto it, we introduce two schedulers, Dynamic and LogFit, which distribute the tasks among all the resources in an optimal manner. A new platform is created based on interrupts that removes spin-locks and allows the processing cores to sleep when not performing useful work. For a compute-intensive application, we obtained up to 45.56% more throughput and 17.89% less energy consumption when all devices of a Zynq-7000 SoC collaborate in the computation compared against FPGA-only execution.

The presentation of Table 2 was incorrect in the original article. The correct Table 2 is given below. The original article has been corr

Fully binarised convolutional neural networks (CNNs) deliver very high inference performance using single-bit weights and activations, together with XNOR type operators for the kernel convolutions. Current research shows that full binarisation results in a degradation of accuracy and different approaches to tackle this issue are being investigated such as using more complex models as accuracy reduces. This paper proposes an alternative based on a multi-precision CNN frame-work that combines a binarised and a floating point CNN in a pipeline configuration deployed on heterogeneous hardware. The binarised CNN is mapped onto an FPGA device and used to perform inference over the whole input set while the floating point network is mapped onto a CPU device and performs re-inference only when the classification confidence level is low. A light-weight confidence mechanism enables a flexible trade-off between accuracy and throughput. To demonstrate the concept, we choose a Zynq 7020 device as the hardware target and show that the multi-precision network is able to increase the BNN accuracy from 78.5% to 82.5% and the CPU inference speed from 29.68 to 90.82 images/sec.

In heterogeneous computing, efficient parallelism can be obtained if every device runs the same task on a different portion of the data set. This requires designing a scheduler which assigns data chunks to compute units proportional to their throughputs. For FPGA-CPU heterogeneous devices, to provide the best possible overall throughput, a scheduler should accurately evaluate the different performance behaviour of the compute devices. In this article, we propose a scheduler which initially detects the highest throughput each device can obtain for a specific application with negligible overhead and then partitions the dataset for improved performance. To demonstrate the efficiency of this method, we choose a Zynq UltraScale+ ZCU102 device as the hardware target and parallelise four applications showing that the developed scheduler can provide up to 94.06% of the throughput achievable at an ideal condition, with comparable power and energy consumption.

In this article, we investigate how to utilise an Field-Programmable Gate Array (FPGA) in an embedded system to save energy. For this purpose, we study the energy efficiency of a hybrid FPGA-CPU device that can switch task execution between hardware and software with a focus on periodic tasks. To increase the applicability of this task switching, we also consider the voltage and frequency scaling (VFS) applied to the FPGA to reduce the system energy consumption. We show that in some cases, if the task’s period is higher than a specific level, the FPGA accelerator cannot reduce the energy consumption associated to the task and the software version is the most energy efficient option. We have applied the proposed techniques to a robot map creation algorithm as a case study which shows up to 38% energy reduction compared to the FPGA implementation. Overall, experimental results show up to 48% energy reduction by applying the proposed techniques at runtime on 13 individual tasks.

This paper proposes a systematic approach to help designers to optimise a given streaming application for FPGAs using High-Level Synthesis (HLS). The proposed technique specifically addresses the two main issues in a streaming application that are determining the exact amount of loop unrolling in the HLS code to increase the throughput and finding the optimum buffers' size to prevent deadlocks. To evaluate the proposed techniques two applications from the machine learning optimisation area are studied in the paper. These applications are Hessian-vector product and Conjugate Gradient (CG). The experimental results show up to 38× speed-up in throughput compared to the original streaming implementations provided by knowledgeable engineers using the dataflow, loop pipelining and FIFO channel related pragmas provided by the HLS tool. In addition, these applications show up to 2.98 GB/sec usage of memory bandwidth which is 93.1% of the total memory bandwidth available on the system. The source codes of the designs are available at https://github.com/Hosseinabady/csdfg-hls.

This paper investigates the energy reductions possible in commercially available FPGAs configured to support voltage, frequency and logic scalability combined with power gating. Voltage and frequency scaling is based on in-situ detectors that allow the device to detect valid working voltage and frequency pairs at run-time while logic scalability is achieved with partial dynamic reconfiguration. The considered devices are FPGA-processor hybrids with independent power domains fabricated in 28 nm process nodes. The test case is based on a number of operational scenarios in which the FPGA side is loaded with a motion estimation core that can be configured with a variable number of execution units. The results demonstrate that voltage scalability reduces power by up to 60 percent compared with nominal voltage operation at the same frequency. The energy analysis show that the most energy efficiency core configuration depends on the performance requirements. A low performance scenario shows that serial computation is more energy efficient than the parallel configuration while the opposite is true when the performance requirements increase. An algorithm is proposed to combine effectively adaptive voltage/logic scaling and power gating in the proposed system and application.

In this paper, we propose a technique to improve the energy efficiency of FPGA devices by exploiting power gating during idle periods in streaming applications. The main idea is to shuffle idle periods during application execution so that the energy and timing overheads of turning the FPGA on and off can become acceptable. A key requirement is that fast FPGA-based accelerators are available and that the application follows a repetitive nature of execution. In this case, the accelerators work on a successive computing mode to accumulate the idle intervals in different iterations in order to make power gating feasible. Streaming on demand applications which are ubiquitous in embedded and portable devices are very good candidates to benefit from this technique. A case study is presented based on an MP3 player as the streaming application which shows up to 52.9% energy reduction.

This paper presents a workgroup synthesis mechanism to compile an OpenCL kernel to FPGA-based accelerators embedded in a multi-core CPU system-on-a-chip (SoC). The OpenCL kernels considered in this paper exhibit regular data access patterns. Coping with the limited amount of internal memory in embedded FPGAs, the workgroup synthesis utilises a novel data access pattern formulation to describe the parallelism already provided by the OpenCL kernels. To provide an OpenCL framework prototype to validate the proposed technique, a source-to-source compiler that transforms the OpenCL kernel into C/C++ code is developed. Then vendor-specific high-level synthesis tools are used to convert the C/C++ code into the FPGA bitstream. Results based on popular real applications show up to 89.8% improvement in the execution time compared to other commercial FPGA OpenCL implementations.

Energy proportional computing (EPC) enables the allocation of energy to tasks depending on computational demands. Computing at full speed and then dynamically turning off modules when they are not required for a period of time can be used to obtain EPC and it is an alternative to voltage scaling techniques in which the computation is slowed down. This paper investigates the viability of physical power gating FPGA devices that incorporate a hardened processor in a different power domain. The run-time power gating approach is applied to Xilinx ZYNQ devices that incorporate a hardened Cortex A9 multi-processor. The paper demonstrates that power down followed by a full reconfiguration can be controlled by the embedded processor autonomously. The results show that the minimum time that the FPGA fabric must remain in power-off state for the technique to be energy efficient is in the order of milliseconds and up to 96% power reduction occurs when the fabric voltage is lowered below critical level. These results take into account the overheads of controlling the programmable voltage regulators interfaced to the FPGA and the overhead of the reconfiguration needed when the device must be returned to the active state.

Employing thousands of cores in a single chip is the natural trend to handle the ever increasing performance requirements of complex applications such as those used in graphics and multimedia processing. System-on-chips (SoCs) platforms based on network-on-chips (NoCs) could be a viable option for the deployment of large multicore designs with thousands of cores. This paper proposes the generalized binary de Bruijn (GBDB) graph as a reliable and efficient network topology for a large NoC. We propose a reliable routing algorithm to detour a faulty channel between two adjacent switches. In addition, using integer linear programming, we propose an optimal tile-based implementation for a GBDB-based NoC in which the number of channels is less than that of Torus which has the same number of links. Our experimental results show that the latency and energy consumption of the generalized de Bruijn graph are much less than those of Mesh and Torus. The low energy consumption of a de Bruijn graph-based NoC makes it suitable for portable devices which have to operate on limited batteries. Also, the gate level implementation of the proposed reliable routing shows small area, power, and timing overheads due to the proposed reliable routing algorithm.

The IEEE SystemC standard has been accepted as an effective high-level system modelling library among designers. However, in order to implement fast simulation models and to consider new ideas and requirements at system level, some enhancements and new features should be added to this standard. This is the reason why OSCI has proposed the TLM 1&2 libraries. In this work, we investigate a very fast but accurate and simple to use modelling methodology for systems which include a large number of modules such as complex Network-on-Chips (NoCs) which have many routers, network interfaces and processing cores. The proposed methodology implements all SystemC processes using normal functions and supports process activation simply by function calls. For this purpose, it utilises a SystemC method process that establishes concurrency and communication among the functions. The experimental results show an improvement of up to 98% in elaboration time and up to 90% in simulation time for small size NoCs. In addition, it can efficiently simulate large NoC with tens of thousands of nodes while traditional SystemC modelling based on threads struggles to simulate hundreds of nodes.

This paper investigates the challenges of run-time resource management in future coarse-grained network-on-reconfigurable-chips (NoRCs). Run-time reconfiguration is a key feature expected in future processing systems which must support multiple applications whose processing requirements are not known at design time. This paper investigates a stochastic routing algorithm in a NoC-based system with dynamically reconfigurable tiles, able to cope with the dynamic behaviour of run-time task mapping. Experimental results show the efficiency of the proposed stochastic task mapping.

This paper proposes a network-on-chip (NoC)-based dynamically reconfigurable platform which can perform multiple applications, simultaneously. A tile attached to a router in the NoC consists of a core container which can host a core permanently or temporarily. The tile also has a hardwired controller and a cache like memory to control the hosted cores. A core, which runs a task, may be described by a bitstream (called hardware core) or a programme code (called software core). Because of the dynamic behaviour of the proposed platform, using task identifier, a stochastic dynamic routing algorithm will find (or map) the task in the platform. Because of using the task identifier in routing algorithm and the reconfigurability of tiles, the proposed platform can tolerate probable faults. The proposed SoC architecture is easily able to run new protocols and tasks. Our results show that, the proposed platform follows the user interests such that runs tasks with higher temporal locality much faster than the tasks with lower temporal locality.

In this paper, we use the generalized binary de Bruijn (GBDB) graph as a scalable and efficient network topology for an on-chip communication network. Using just two-layer wiring, we propose an optimum tile-based implementation for a GBDB- based Network-on-Chip (NoC). Our experimental results show that the latency and energy consumption of generalized de Bruijn graph are much less with compared to Mesh and Torus, the two common NoC architectures in the literature.

Motivated by the problems associated with soft errors in digital circuits and fault related attacks in cryptographic hardware, we presented a systematic method for designing single error correcting multiplier circuits for finite fields or Galois fields over GF(2m) in [7]. We used multiple parity predictions to correct single errors based on the Hamming principles. The problem with Hamming based error correction is the delay overhead. To mitigate the delay overhead, in this paper we present single error correction using Low Density Parity Check Codes (LDPC). The expressions for the parity prediction are derived from the input operands, and are based on the primitive polynomials of the fields. Our technique, when compared with existing techniques, gives better performance. We show that our Single Error Correction (SEC) multipliers over GF(2m) require slightly over 100 percent extra hardware, whereas with the traditional SEC techniques this figure is more than 200 percent.

Time, power, and data volume are among some of the most challenging issues for testing system-on-chip (SoC) and have not been fully resolved, even if a scan-based technique is employed. A novel architecture, referred to the selective trigger scan architecture, is introduced in this paper to address these issues. This architecture reduces switching activity in the circuit-under-test (CUT) and increases the clock frequency of the scanning process. An auxiliary chain is utilized in this architecture to avoid the large number of transitions to the CUT during the scan-in process, as well as enabling retention of the currently applied test vectors and applying only necessary changes to them. The auxiliary chain shifts in the difference between consecutive test vectors and only the required transitions (referred to as trigger data) are applied to the CUT. Power requirements are substantially reduced; moreover, DFT penalties are reduced because no additional multiplexer is utilized along the scan path. Data reformatting is applied in order to make the proposed architecture amenable to data compression, thus permitting a further reduction in test time. It also permits delay fault testing. Using ISCAS 85 and 89 benchmark circuits, the effectiveness of this architecture for improving SoC test measures (such as power, time, and data volume) is experimentally evaluated and confirmed.

This paper proposes an efficient test methodology to test switches in a network-on-chip (NoC) architecture. A switch in a NoC consists of a number of ports and a router. Using the intra-switch regularity among ports of a switch and inter-switch regularity among routers of switches, the proposed method decreases the test application time and test data volume of NoC testing. Using a test source to generate test vectors and scan-based testing, this methodology broadcasts test vectors through the minimum spanning tree of the NoC and concurrently tests its switches. In addition, a possible fault is detected by comparing test results using inter- or intra- switch comparisons. The logic and memory parts of a switch are tested by appropriate memory and logic testing methods. Experimental results show less test application time and test power consumption, as compared with other methods in the literature

This paper proposes an analytical model to assess Reliability Factor of an NoC based System-on-Chip design. Reliability Factor is the probability that faults in the NoC infrastructure can be recovered without any effect on system functionality. The proposed method classifies switch faults of an NoC according to their impact on system functionality. Based on this classification, the contribution of each transient fault lowering the reliability of the NoC is calculated. This model can be used to decide which fault tolerant techniques cause more improvement on system reliability.

This paper proposes an analytical method to assess the soft-error rate (SER) in the early stages of a System-on-Chip (SoC) platform-based design methodology. The proposed method gets an executable UML (Unified Modeling Language) model of the SoC and the raw softerror rate of different parts of the platform as its inputs. Soft-errors on the design are modeled by disturbances on the value of attributes in the classes of the UML model and disturbances on opcodes of software cores. The Dynamic behavior of each core is used to determine the propagation probability of each variable disturbance to the core outputs. Furthermore, the SER and the execution time of each core in the SoC and a Failure Modes and Effects Analysis (FMEA) that determines the severity of each failure mode in the SoC are used to compute the System-Failure Rate (SFR) of the SoC.

This paper proposes an analytical method to assess soft-error rate (SER) in the early stages of a system-on-chip (SoC) platform-based design methodology. The proposed method uses an executable UML model of the SoC for its input. Soft-errors on the design are modeled by disturbances on the value of attributes in the classes of the UML model and disturbances on opcodes of software cores. SER and execution time of each core in the SoC and a failure modes and effects analysis (FMEA) that determines the severity of each failure mode in the SoC are used to compute the system-failure rate (SFR) of the SoC

Recent advances in process technology have led to a rapid increase in the density of integrated circuits (ICs). Increased density and the need to test for new types of defects in nanometer technologies have resulted in a tremendous increase in test application time (TAT). This article presents a test synthesis method to reduce test application time for testing the datapath of a design. The test application time is reduced by applying a test-time-aware resource sharing algorithm on a scheduled control data flow graph (CDFG) of a design.

This paper proposes reuse of on-chip networks for testing switches in network on chips (NoCs). The proposed algorithm broadcasts test vectors of switches through the on-chip networks and detects faults by comparing output responses of switches with each other. This algorithm alleviates the need for: (1) external comparison of the output response of the circuit-under-test with the response of a fault free circuit stored on a tester (2) on-chip signature analysis (3) a dedicated test-bus to reach test vectors and collect their responses. Experimental results on a few test benches compare the proposed algorithm with traditional system on chip (SoC) test methods

The effect of single-event transients (SETs) (at a combinational node of a design) on the system reliability is becoming a big concern for ICs manufactured using advanced technologies. An SET at a node of combinational part may cause a transient pulse at the input of a flip-flop and consequently is latched in the flip-flop and generates a soft-error. When an SET conjoined with a transition at a node along a critical path of the combinational part of a design, a transient delay fault may occur at the input of a flip-flop. On the other hand, increasing pipeline depth and using low power techniques such as multi-level power supply, and multi-threshold transistor convert almost all paths in a circuit to critical ones. Thus, studying the behavior of the SET in these kinds of circuits needs special attention. This paper studies the dynamic behavior of a circuit with massive critical paths in the presence of an SET. We also propose a novel flip-flop architecture to mitigate the effects of such SETs in combinational circuits. Furthermore, the proposed architecture can tolerant a single event upset (SEU) caused by particle strike on the internal nodes of a flip-flop

Formal verification of microprocessors requires a mechanism for efficient representation and manipulation of both arithmetic and random Boolean functions. Recently, a new canonical and graph-based representation called TED has been introduced for verification of digital systems. Although TED can be used effectively to represent arithmetic expressions at the word-level, it is not memory efficient in representing bit-level logic expressions. In this paper, we present modifications to TED to improve its ability for bit-level logic representation while maintaining its robustness in arithmetic word-level representation. It will be shown that for random Boolean expressions, the modified TED performs the same as BDD representation.

Power dissipation during test is a major challenge in testing integrated circuits. Dynamic power has been the dominant part of power dissipation in CMOS circuits, however, in future technologies the static portion of power dissipation will outreach the dynamic portion. This paper proposes an efficient technique to reduce both dynamic and static power dissipation in scan structures. Scan cell outputs which are not on the critical path(s) are multiplexed to fixed values during scan mode. These constant values and primary inputs are selected such that the transitions occurring on nonmultiplexed scan cells are suppressed and the leakage current during scan mode is decreased. A method for finding these vectors is also proposed. The effectiveness of this technique is proved by experiments performed on ISCAS89 benchmark circuits.