三种高速乘法器的实现原理及性能比较

本文介绍了三种高速乘法器架构：阵列乘法器、修正布斯算法(MBA)乘法器、华莱士(WT)乘法器，并对基于以上三种架构的32位乘法器性能进行了比较。选择乘法器，应根据实际应用。从面积、速度、功耗等角度权衡考虑。     

可配置GF(2m)域Digit-Serial乘法器

本文针对椭圆加密算法的应用，基于已有的GF(2^m)域Digit—Serial不可配置乘法器，通过控制输入数据格式、内镶GF（2^m）域Digit—Serial不可配置乘法器，得到了一个在硬件上可配置的快速乘法器。运用本文的思想实现了可计算域值为150～256的GF(2^m)域Digit-Serial的乘法器，用此乘法器计算域值为163的乘法，仿真结果同域值为163的不可配置并行乘法器的一致。本文最后还给出了几种可配置乘法器结构的性能比较，结果表明在硬件上可配置的GF（2^m）域乘法器解决方案中，本文提出的结构克服了并行可配置乘法器在大域值应用中关键路径延迟太长、硬件开销太大，串行可配置乘法器实现速度太慢的弊病。需要说明的是，本文的实现方法可以内镶各种不同的GF（2^m）域Digit-Serial不可配置乘法器以满足实际应用的需要。     

GF（2^8）上快速乘法器及求逆器的设计

基于多项式乘法理论，采用高层次设计方法，设计并采用ＦＰＧＡ实现了ＧＦ（２＾８）上８位快速乘法器，并利用该乘法器设计了一个计算ＧＦ（２＾８）上任一元素的例数的求逆器，该乘法器与求逆器可以应用于ＲＳ（２５５．２２３）码编／译码器。     

高速浮点乘法器设计

设计了一种符合IEEE-754标准的32bits高速CMOS浮点乘法器.该乘法器采用MBA算法和基于4:2 compressor的树型结构完成Carry Save形式的部分积压缩,再由高速Carry Select加法器求得乘积.电路设计采用了新型的高速加法运算单元.乘法器采用0.35(m制程,内含19,197个晶体管.3.3V工作电压下(室温),乘法器延迟时间为3.807ns,功耗为107mW@100MHz.     

WT及其在SR中的应用

小波变换（WT：Wavelet）为处理语音、音乐等时变信号提供了一个新的方法，可以通过选取满足一定条件的Wavelet波形来表征人类听觉系统对声音信号的感知，本文对WT和短时FT（STFT）作了比较，并讨论了WT在语音识别（SR）中的应用，最后给出了实验结果。     

FDM调制解调中高速（I）FFT硬件设计

提出了一种64点，512点和1024点（I）FFT（（逆）快速傅里叶变换）的硬件实现方法，适合应用在正交频分复用（OFDM）系统中，实现时采用了16位精度的复数来表示输入输出数据。该算法在运算过程去除了所有的乘法器在运算过程中没有使用乘法器，使得运算速度得到较大地提高。     

待机功耗测量软件

该功耗测量软件符合“IEC62301 Ed1．0（2005）标准和European EuP Directive”规定的测试方法,可免费下载,适用于横河WT3000、WT1800、WT1600、WT500和WT210分析仪,此软件可使用户依据“IEC62301Ed1．0（2005）：家用电器一待机功率测量和EuP（能源使用产品）Directive Lot6”标准进行测量。     

光存储RS译码器设计与译码速度分析

比较了reed-solomon（RS）译码的Berlekamp-Massey（BM）算法和Euclidean算法的运行速度，并选择BM算法设计了满足36Mbps数据传输率（D豫）的RS译码器。针对现有几种光盘的DTR，进一步分析了光存储中RS译码速度的要求，并对译码中的有限域乘法器做了仿真。该乘法器在工作频率为50MHz的FPGA芯片中工作正常，可以满足光盘的DTR要求。     

模拟乘法器MC1496的应用研究

模拟乘法器是一种完成两个模拟信号相乘的电子器件。本文基于电路设计与仿真软件Multisim进行模拟乘法器MC1496的应用研究,具体实现普通调幅（AM）、双边带调幅（DSB）、同步检波器、混频器的电路设计与仿真。     

高速Booth编码模（2^n—1）乘法器的设计

在余数系统中（2^n-1）是最普遍应用的模,提出了一种新的booth编码结构,并基于提出的booth编码结构,提出了一种高速模（2^n-1）乘法器．该乘法器采用CSA或者wallace Tree结构可以进一步提高运算速度．此乘法器在一个时钟周期内可以完成所需运算,简单高效．     

High speed multiplier using Nikhilam Sutra algorithm of Vedic mathematics

This article presents the design of a new high-speed multiplier architecture using Nikhilam Sutra of Vedic mathematics. The proposed multiplier architecture finds out the compliment of the large operand from its nearest base to perform the multiplication. The multiplication of two large operands is reduced to the multiplication of their compliments and addition. It is more efficient when the magnitudes of both operands are more than half of their maximum values. The carry save adder in the multiplier architecture increases the speed of addition of partial products. The multiplier circuit is synthesised and simulated using Xilinx ISE 10.1 software and implemented on Spartan 2 FPGA device XC2S30-5pq208. The output parameters such as propagation delay and device utilisation are calculated from synthesis results. The performance evaluation results in terms of speed and device utilisation are compared with earlier multiplier architecture. The proposed design has speed improvements compared to multiplier architecture presented in the literature.     

32kb／sADPCM中高速乘法器的设计

本文介绍了６０路３２ｋｂ／ｓＡＤＰＣＭ专用芯片中的高速乘法器的逻辑设计和提高运算速度的方法。通过优化设计，该乘法器运算速度高，电路简单，对芯片制造工艺要求不高。     

Low power and high speed multiplier design with row bypassing and parallel architecture

This paper presents a low power and high speed row bypassing multiplier. The primary power reductions are obtained by tuning off MOS components through multiplexers when the operands of multiplier are zero. Analysis of the conventional DSP applications shows that the average of zero input of operand in multiplier is 73.8 percent. Therefore, significant power consumption can be reduced by the proposed bypassing multiplier. The proposed multiplier adopts ripple-carry adder with fewer additional hardware components. In addition, the proposed bypassing architecture can enhance operating speed by the additional parallel architecture to shorten the delay time of the proposed multiplier. Both unsigned and signed operands of multiplier are developed. Post-layout simulations are performed with standard TSMC 0.18 μm CMOS technology and 1.8 V supply voltage by Cadence Spectre simulation tools. Simulation results show that the proposed design can reduce power consumption and operating speed compared to those of counterparts. For a 16×16 multiplier, the proposed design achieves 17 and 36 percent reduction in power consumption and delay, respectively, at the cost of 20 percent increase of chip area in comparison with those of conventional array multipliers. In addition, the proposed design achieves averages of 11 and 38 percent reduction in power consumption and delay with 46 percent less chip area in comparison with those counterparts for both unsigned and signed multipliers. The proposed design is suitable for low power and high speed arithmetic applications.     

一种规整高效的缩1码模2^n＋1乘法器的VLSI设计

针对目前缩1码模2n＋1乘法器的优缺点,设计出一个有效的缩1码模2n＋1乘法器。该模乘法器是由改进的基-4 Booth编码模块、规整的缩1码进位保留加法器树以及缩1码模加法器构成,部分积的个数减少到n/2＋2个,具有统一的编码电路,简单的校正项生成电路,较快的计算速度,尤其是能够处理操作数和结果为0的情况,实现了操作数的全输入。比较结果表明,该模乘法器在同类型模乘法器中以最少的面积获得了更快的速度。     

一种新型的高速FIR滤波器及其VLSI实现

本文提出了一种新型的高速滤波器结构,此结构的核心是一种独特的乘加单元.该乘加单元是通过对BOOTH型乘法器与高速加法器结构的深入研究而探索出来的.采用该乘加单元我们可以实现任何阶数高速FIR滤波器.在文章的最后我们采用该结构实现了视频编码器中的一个高速色度滤波器,并与采用传统结构设计的该滤波器进行了性能比较.     

Reduced computational redundancy implementation of DSP algorithms using computation sharing vector scaling

In this paper, we present a general approach which specifically targets reduction of redundant computation in common digital-signal processing (DSP) tasks such as filtering and matrix multiplication. We show that such tasks can be expressed as multiplication of vectors by scalars and this allows fast multiplication by sharing computation. Vector scaling operation is decomposed to find the most effective precomputations which yield a fast multiplier implementation. Two decomposition approaches are presented, one based on a greedy decomposition and the other based on fixed-size lookup and this leads to two multiplier architectures for vector-scalar products. Analog simulation of an example multiplier shows a speed advantage by a factor of about 1.85 over a conventional carry save array multiplier. Further simulations using 0.18 /spl mu/ technology show up to 20% speed advantage over Booth encoded Wallace tree multipliers.     

用EMODL实现的高速低功耗流水线乘法器

实现快速、低功耗以及节省面积的乘法器对高性能微处理器 (例如 DSP和 RISC)而言是至关重要的。文中详尽论述了新型的增强型多输出多米诺逻辑 ( EMODL)及其 n-MOS赋值树的尺寸优化方法 ,并用它实现了高速低功耗 2 0× 2 0 bit流水线乘法器。最后 ,通过 HSPICE仿真 ,确认了该乘法器结构的优越性 :流水线等待时间小 ( 2倍于系统时钟 )、运算速度高 ( 10 0 MOPS)以及低功耗 ( 2 3 .94m W)     

高速流水线结构的大整数乘法器FPGA设计与实现

大整数乘法是密态数据计算中最为耗时的基本运算操作,提高大数乘法单元的计算速度在全同态加密机器学习等应用中尤为重要.提出了一种输入数据位宽为768 kbit的高速大整数乘法器设计方案,将核心组件64 k点有限域快速数论变换(NTT)分解成16点NTT实现,并通过算法分治处理,细化16点NTT的流水线处理过程.采用加法和移...     

基于优化电路的高性能乘法器设计

为了提高二进制乘法器的速度并降低其功耗,在乘法器的部分积产生模块采用了改进的基4Booth编码和部分积产生电路并在部分积压缩模块应用了7∶3压缩器电路,设计并实现了一种高性能的33×28二进制乘法器.在TSMC 90 nm工艺和0.9 V工作电压下,仿真结果与Synopsys公司module compiler生成的乘法器相比,部分积产生电路速度提高34%,7∶3压缩器和其他压缩器的结合使用减少了约一级全加器的延时,整体乘法器速度提高约17.7%.     

Low energy and area efficient quaternary multiplier with carbon nanotube field effect transistors

In this study, new multiplier and adder method designs with multiplexers are proposed. The designs are based on quaternary logic and a carbon nanotube field-effect transistor (CNTFET). The design utilizes 4 × 4 multiplier blocks. Applying specific rotational functions and unary operators to the quaternary logic reduced the power delay produced (PDP) circuit by 54% and 17.5% in the CNTFETs used in the adder block and by 98.4% and 43.62% in the transistors in the multiplier block, respectively. The proposed 4 × 4 multiplier also reduced the occupied area by 66.05% and increased the speed circuit by 55.59%. The proposed designs are simulated using HSPICE software and 32 nm technology in the Stanford Compact SPICE model for CNTFETs. The simulated results display a significant improvement in the fabrication, average power consumption, speed, and PDP compared to the current best-performing techniques in the literature. The proposed operators and circuits are evaluated under various operating conditions, and the results demonstrate the stability of the proposed circuits.