Streaming Elements for FPGA Signal and Image Processing Accelerators

Field programmable gate array devices boast abundant resources with which custom accelerator components for signal, image and data processing may be realised; however, realising high performance, low cost accelerators currently demands manual register transfer level design. Software-programmable ’soft’ processors have been proposed as a way to reduce this design burden but they are unable to support performance and cost comparable to custom circuits. This paper proposes a new soft processing approach for FPGA which promises to overcome this barrier. A high performance, fine-grained streaming processor, known as a Streaming Accelerator Element, is proposed which realises accelerators as large scale custom multicore networks. By adopting a streaming execution approach with advanced program control and memory addressing capabilities, typical program inefficiencies can be almost completely eliminated to enable performance and cost which are unprecedented amongst software-programmable solutions. When used to realise accelerators for fast fourier transform, motion estimation, matrix multiplication and sobel edge detection it is shown how the proposed architecture enables real-time performance and with performance and cost comparable with hand-crafted custom circuit accelerators and up to two orders of magnitude beyond existing soft processors.

Energy versus data integrity trade-offs in embedded high-density logic compatible dynamic memories

Current variation aware design methodologies, tuned for worst-case scenarios, are becoming increasingly pessimistic from the perspective of power and performance. A good example of such pessimism is setting the refresh rate of DRAMs according to the worst-case access statistics, thereby resulting in very frequent refresh cycles, which are responsible for the majority of the standby power consumption of these memories. However, such a high refresh rate may not be required, either due to extremely low probability of the actual occurrence of such a worst-case, or due to the inherent error resilient nature of many applications that can tolerate a certain number of potential failures. In this paper, we exploit and quantify the possibilities that exist in dynamic memory design by shifting to the so-called approximate computing paradigm in order to save power and enhance yield at no cost. The statistical characteristics of the retention time in dynamic memories were revealed by studying a fabricated 2kb CMOS compatible embedded DRAM (eDRAM) memory array based on gain-cells. Measurements show that up to 73% of the retention power can be saved by altering the refresh time and setting it such that a small number of failures is allowed. We show that these savings can be further increased by utilizing known circuit techniques, such as body biasing, which can help, not only in extending, but also in preferably shaping the retention time distribution. Our approach is one of the first attempts to access the data integrity and energy tradeoffs achieved in eDRAMs for utilizing them in error resilient applications and can prove helpful in the anticipated shift to approximate computing.

Signal Reliability Improvement Using Selection Combining Based Macro-Diversity for Off-Body Communications At 868 MHz

This paper investigates the potential improvement in signal reliability for outdoor short-range off-body communications channels at 868 MHz using the macro-diversity offered by multiple co-located base stations. In this study, ten identical hypothetical base stations were positioned equidistantly around the perimeter of a rectangle of length 6.67 m and width 3.3 m. A body worn node was placed on the central chest region of an adult male. Five scenarios, each considering different user trajectories, were then analyzed to test the efficacy of using macro-diversity when the desired link is subject to shadowing caused by the human body. A number of selection combining based macro-diversity configurations consisting of four and then ten base stations were considered. It was found that using a macro-diversity system consisting of four base stations (or equivalently signal branches), a maximum diversity gain of 22.5 dB could be obtained while implementing a 10-base station setup this figure could be improved to 25.2 dB.

Signal Reception Characteristics in the Proximity of Alice and Bob for Secure Indoor Peer-to-Peer Communications At 2.45 GHz

Mutual variation of the received signal which occurs as a consequence of the channel reciprocity property has recently been proposed as a viable method for secret key generation. However, this cannot be strictly maintained in practice as the property is applicable only in the absence of interference. To ensure the propagation defined key remains secret, one requirement is that there remain high degrees of uncertainty between the legitimate users channel response and that of any eavesdropper's. In this paper, we investigate whether such de-correlation occurs for an indoor point-to-point link at 2.45 GHz. This is achieved by computing the localized correlation coefficient between the simultaneous channel response measured by the legitimate users and that of multiple distributed eavesdroppers for static and dynamic scenarios.

Real-Time Channel Model Selection Using Windowed Received Signal Strength Measurements

This paper investigates the potential for using the windowed variance of the received signal strength to select from a set of predetermined channel models for a wireless ranging or localization system. An 868 MHz based measurement system was used to characterize the received signal strength (RSS) of the off-body link formed between two wireless nodes attached to either side of a human thorax and six base stations situated in the local surroundings.

DRHA 2015: Bringing practice closer to research - seeking integrity, sincerity and authenticity

A paper on 'practice as research' (or as I emphasise in this presentation as "Practice Research") debating current dialogue around the issue of practice and research.

Specific role of neuronal nitric-oxide synthase when tethered to the plasma membrane calcium pump in regulating the beta-adrenergic signal in the myocardium

The cardiac neuronal nitric-oxide synthase (nNOS) has been described as a modulator of cardiac contractility. We have demonstrated previously that isoform 4b of the sarcolemmal calcium pump (PMCA4b) binds to nNOS in the heart and that this complex regulates beta-adrenergic signal transmission in vivo. Here, we investigated whether the nNOS-PMCA4b complex serves as a specific signaling modulator in the heart. PMCA4b transgenic mice (PMCA4b-TG) showed a significant reduction in nNOS and total NOS activities as well as in cGMP levels in the heart compared with their wild type (WT) littermates. In contrast, PMCA4b-TG hearts showed an elevation in cAMP levels compared with the WT. Adult cardiomyocytes isolated from PMCA4b-TG mice demonstrated a 3-fold increase in Ser(16) phospholamban (PLB) phosphorylation as well as Ser(22) and Ser(23) cardiac troponin I (cTnI) phosphorylation at base line compared with the WT. In addition, the relative induction of PLB phosphorylation and cTnI phosphorylation following isoproterenol treatment was severely reduced in PMCA4b-TG myocytes, explaining the blunted physiological response to the beta-adrenergic stimulation. In keeping with the data from the transgenic animals, neonatal rat cardiomyocytes overexpressing PMCA4b showed a significant reduction in nitric oxide and cGMP levels. This was accompanied by an increase in cAMP levels, which led to an increase in both PLB and cTnI phosphorylation at base line. Elevated cAMP levels were likely due to the modulation of cardiac phosphodiesterase, which determined the balance between cGMP and cAMP following PMCA4b overexpression. In conclusion, these results showed that the nNOS-PMCA4b complex regulates contractility via cAMP and phosphorylation of both PLB and cTnI.

A BRCA1 deficient, NFκB driven immune signal predicts good outcome in triple negative breast cancer

Triple negative (TNBCs) and the closely related Basal-like (BLBCs) breast cancers are a loosely defined collection of cancers with poor clinical outcomes. Both show strong similarities with BRCA1-mutant breast cancers and BRCA1 dysfunction, or 'BRCAness', is observed in a large proportion of sporadic BLBCs. BRCA1 expression and function has been shown in vitro to modulate responses to radiation and chemotherapy. Exploitation of this knowledge in the treatment of BRCA1-mutant patients has had varying degrees of success. This reflects the significant problem of accurately detecting those patients with BRCA1 dysfunction. Moreover, not all BRCA1 mutations/loss of function result in the same histology/pathology or indeed have similar effects in modulating therapeutic responses. Given the poor clinical outcomes and lack of targeted therapy for these subtypes, a better understanding of the biology underlying these diseases is required in order to develop novel therapeutic strategies.We have discovered a consistent NFκB hyperactivity associated with BRCA1 dysfunction as a consequence of increased Reactive Oxygen Species (ROS). This biology is found in a subset of BRCA1-mutant and triple negative breast cancer cases and confers good outcome. The increased NFκB signalling results in an anti-tumour microenvironment which may allow CD8+ cytotoxic T cells to suppress tumour progression. However, tumours lacking this NFκB-driven biology have a more tumour-promoting environment and so are associated with poorer prognosis. Tumour-derived gene expression data and cell line models imply that these tumours may benefit from alternative treatment strategies such as reprogramming the microenvironment and targeting the IGF and AR signalling pathways.

Inexact-aware architecture design for ultra-low power bio-signal analysis

This study introduces an inexact, but ultra-low power, computing architecture devoted to the embedded analysis of bio-signals. The platform operates at extremely low voltage supply levels to minimise energy consumption. In this scenario, the reliability of static RAM (SRAM) memories cannot be guaranteed when using conventional 6-transistor implementations. While error correction codes and dedicated SRAM implementations can ensure correct operations in this near-threshold regime, they incur in significant area and energy overheads, and should therefore be employed judiciously. Herein, the authors propose a novel scheme to design inexact computing architectures that selectively protects memory regions based on their significance, i.e. their impact on the end-to-end quality of service, as dictated by the bio-signal application characteristics. The authors illustrate their scheme on an industrial benchmark application performing the power spectrum analysis of electrocardiograms. Experimental evidence showcases that a significance-based memory protection approach leads to a small degradation in the output quality with respect to an exact implementation, while resulting in substantial energy gains, both in the memory and the processing subsystem.