This article proposes an open-source hardware Posit Arithmetic Core Generator (PACoGen) for the recently developed universal number posit number system, along with a set of pipelined architectures. Posit number system composed of a run-time varying exponent component, which is defined by a composition of varying length “regime-bit„ and “exponent-bit„ (with a maximum size of ES bits, the exponent size). This in effect also makes the fraction part to vary at run-time in size and position. These run-time variations inherits an interesting hardware design challenge for posit arithmetic architectures. Posit number system, being at an infant stage of its development, possess very limited hardware solutions for its arithmetic architectures. In this view, this paper targets the algorithmic development and generic HDL generators (PACoGen) for basic posit arithmetic. The proposed open source PACoGen currently includes the adder/subtractor, multiplier and division arithmetic. PACoGen can provide the Verilog HDL code respective posit arithmetic for any given posit word-width (N) and exponent size (ES), as defined under the posit number system. Further, pipelined architectures of 32-bit posit with 6-bit exponent size are also proposed and discussed for addition/subtraction, multiplication and division arithmetic. The proposed posit arithmetic architectures are demonstrated on Virtex-7 (xc7vx330t-3ffg1157) FPGA device as well as Nangate 15nm ASIC platform. PACoGen would open a gateway for further posit arithmetic hardware exploration and evaluation.
13 October 2016 in IEEE Transactions on Circuits and Systems I: Regular Papers
Floating point division is a core arithmetic widely used in scientific and engineering applications. This paper proposed an architecture for double precision floating point division. This architecture is designed for dual-mode functionality, which can either compute on a pair of double precision operands or on two pairs of single precision operands in parallel. The architecture is based on the series expansion multiplicative methodology of mantissa computation. For this, a novel dual-mode Radix-4 Modified Booth multiplier is designed, which is used iteratively in the architecture of dual-mode mantissa computation. Other key components of floating point division flow (such as leading-one-detection, left/right dynamic shifters, rounding, etc.) are also re-designed for the dual-mode operation. The proposed dual-mode architecture is synthesized using UMC 90 nm technology ASIC implementation. Two versions of proposed architecture are presented, one with single stage multiplier and another with two stage multiplier. Compared to a standalone double precision division architecture, the proposed dual-mode architecture requires 17% to 19% extra hardware resources, with 3% to 5% period overhead. In comparison to prior art on this, the proposed architecture out-performs them in terms of required area, time-period and throughput.
Large integer multiplication and floating point multiplication are the two dominating operations for many scientific and cryptographic applications. Large integer multipliers generally have linearly but high area requirement according to a given bit-width. High precision requirements of a given application lead to the use of quadruple precision arithmetic, however its operation is dominated by large integer multiplication of the mantissa product. In this paper, we propose a hardware efficient approach for implementing a fully pipelined large integer multipliers, and further extending it to Quadruple Precision (QP) floating point multiplication. The proposed design uses less hardware resources in terms of DSP48 blocks and slices, while attaining high performance. Promising results are obtained when compared our designs with the best reported large integer multipliers and also QP floating point multiplier in literatures. For instance, our results have demonstrated a significant improvement for the proposed QP multiplier, for over 50% improvement in terms of the DSP48 block usage with a penalty of slight additional slices, when compared to the best result in the literature on a Virtex-4 device.
Floating point multiplication is a crucial and useful arithmetic operation for many scientific and signal processing applications. High precision requirements of many applications lead to the incorporation of quadruple precision (QP) arithmetics. The logic complexity and performance overhead of quadruple precision arithmetic are quite large. This paper has focused on one of the quadruple precision arithmetic operations, multiplication. We present an efficient implementation of QP multiplication operation on a reconfigurable FPGA platform. The presented design uses much less hardware resource in terms of DSP48 blocks, and slices with a higher performance. Promising results are obtained by comparing the proposed designs with the best reported QP floating point multiplier in the literature. We have achieved more than 50% improvements in the amount of DSP48 block at a slight cost of additional slices, on a Virtex-4 FPGA.
17 April 2017 in Circuits, Systems, and Signal Processing