This paper presents a multiple-voltage high-level synthesis approach for low power DSP applications using algorithmic transformation techniques. Our approach is motivated by maximization of task mobilities in that the increase of mobilities may raise the possibility of assigning tasks to low-voltage components. The mobility means the ability to schedule the starting time of a task. It is defined as the distance between its as-late-as-possible (ALAP) schedule time and its as-soon-as-possible (ASAP) schedule time. To earn task mobilities, we use loop shrinking, retiming and unfolding techniques. The loop shrinking can first reduce the iteration period bound (IPB) and, then, the others are employed for shortening the iteration period (IP) as much as possible. The minimization of IP results in high task mobilities. Finally, we can assign tasks with high mobilities to low-voltage components and, thus, minimize energy under resource and latency constraints. With considering the overhead of level conversion, our approach can achieve significant power reduction. In the case of the third-order IIR filter, the proposed approach can save up to 40.2% of power consumption.

This paper presents a memory-efficient motion estimation (ME) technique for high-resolution video compression. The main objective is to reduce the external memory access, especially for limited local memory resource. The reduction of memory access can successfully save the notorious power consumption. The key to reduce the memory accesses is based on center-biased algorithm in that the center-biased algorithm performs the motion vector (MV) searching with the minimum search data. While considering the data reusability, the proposed dual-search-windowing (DSW) approaches use the secondary windowing as an option per searching necessity. By doing so, the loading of search windows can be alleviated and hence reduce the required external memory bandwidth. The proposed techniques can save up to 81% of external memory bandwidth and require only 135 MBytes/sec, while the quality degradation is less than 0.2 dB for 720 p HDTV clips coded at 8 Mbits/sec.

This paper presents an improved architecture of the multistage multibit sigma-delta modulators (ΣΔMs) for wide-band applications. Our approach is based on two resonator topologies, high-Q cascade-of-resonator-with-feedforward (HQCRFF) and low-Q cascade-of-integrator-with-feedforward (LQCIFF). Because of in-band zeros introduced by internal loop filters, the proposed architecture enhances the suppression of the in-band quantization noise at a low OSR. The HQCRFF-based modulator with single-bit quantizer has two modes of operation, modulation and oscillation. When the HQCRFF-based modulator is operating in oscillation mode, the feedback path from the quantizer output to the input summing node is disabled and hence the modulator output is free of the quantization noise terms. Although operating in oscillation mode is not allowed for single-stage ΣΔM, the oscillation of HQCRFF-based modulator can improve dynamic range (DR) of the multistage (MASH) ΣΔM. The key to improving DR is to use HQCRFF-based modulator in the first stage and have the first stage oscillated. When the first stage oscillates, the coarse quantization noise vanishes and hence circuit nonidealities, such as finite op-amp gain and capacitor mismatching, do not cause leakage quantization noise problem. According to theoretical and numerical analysis, the proposed MASH architecture can inherently have wide DR without using additional calibration techniques.

This paper presents a hardware-efficient folding technique for high-order FIR filtering while considering the tradeoff between the number of processing elements and throughput rate. Given the throughput rate, one can always employ the minimum number of processing elements for saving the implementation cost and figure out a folded architecture. However, applying inefficient folding techniques may result in costly switches and registers. Therefore, our work intends to evaluate the efficiency for folding techniques in terms of the number of registers, and the power dissipation of registers. As shown in the estimation results, while comparing with the published folded architectures under the same throughput rate, the proposed folding technique can turn out less power dissipation and low hardware complexity than the others. The proposed design has been implemented using TSMC 0.18 µm 1P6M technology. As seen in the post-layout simulation, our design can meet the requirement of IS-95 WCDMA pulse shaping FIR filter while the power consumption can be as low as 16.66 mW.

This paper presents a VLSI design methodology for the MAC-level DWT/IDWT processor based on a novel limited-resource scheduling algorithm. The r-split Fully-specified Signal Flow Graph (FSFG) of limited-resource FIR filtering has been developed for the scheduling of the MAC-level DWT/IDWT signal processing. Given a set of architecture constraints and DWT parameters, the scheduling algorithm can generate four scheduling matrices that drive the data path to perform the DWT computation. Because the memory for the inter-octave is considered with the register of FIR filter, the memory size is less than the traditional architecture. Besides, based on the limited-resource scheduling algorithm, an automated DWT processor synthesizer has been developed and generates constrained DWT processors in the form of silicon intelligent property (SIP). The DWT SIP can be embedded into a SOC or mapped to program codes for commercial off-the-shelf (COTS) DSP processors with programmable devices. As a result, it has been successfully proven that a variety of DWT SIPs can be efficiently realized by tuning the parameters and applied for signal processing applications.