Multiport Memory and Oating Point Cordic Pipeline in Jacobium Processing Elements
نویسنده
چکیده
| This paper presents the architecture of Jacobium processing elements. The Jacobium is a dataaow processor aiming at high speed execution of algorithms where Cordic operations play a dominant role. To achieve high throughputs the processing elements are equiped with a deep oating point Cordic pipeline, on-chip multiport memory to buuer operands and results, and several high speed communication busses for communication between processing elements and the host. Communication and processing takes place on-nite stream basis. This keeps control and communication overhead low and control logic small. To reduce control overhead even more, the memory is made accessible by abstract instructions from the program controller which do not involve physical addresses. To reduce ow irregularity eeects, data as well as instructions are queued before being consumed , and the pipeline is able to compress bubbles in the pipe. The architecture of Jacobium processing elements is highly parameterized, which enables tailoring their characteristics for speciic Jacobium applications. The objective of the Jacobium project is to develop a scalable processor, the Jacobium, capable of performing high demanding real time calculations appearing in Jacobi algorithms, where Cordic operations play a dominant role 1]. Within the scope of the project, research and development concentrates on: the development of a hardware architectural template for the processing elements in the Jacobium, the development of a design space explorer to evaluate possible conngurations of the Jacobium and its processing elements, and to identify ee-cient conngurations, the development of methods for the mapping of algorithm dependence graphs onto architecture conngurations. The main application that the Jacobium group currently focusses on is digital array processing, an example of which is digital adaptive beamforming. Beam-forming is the process wherein the antenna gain pattern of a phased array antenna is focused in a particular direction. The result is a spatial lter able to receive signals of interest arriving from a particular direction while supressing undesired signals from possibly moving sources from other directions. The application of Cordic-based algorithms in radar signal processing is discussed in 2]. Apart from radar signal processing applications, Cordic based algorithms may prove usefull in other real-time adaptive signal processing applications as well. These other applications may impose diierent requirements on the Jacobium and its processing elements. Therefore the architecture of the processing elements is developed to be highly parameterized. Parameters such as arithmetic accuracy and available on-chip memory resources are easily altered and implemented in new …
منابع مشابه
Efficient implementations of pipelined CORDIC based IIR digital filters using fast orthonormal μ-rotations
CORDIC based IIR digital lters are orthogonal lters whose internal computations consist of orthogonal transformations. These lters possess desirable properties for VLSI implementations such as regularity, local connection, low sensitivity to nite word-length implementation, and elimination of limit cycles. Recently, ne-grain pipelined CORDIC based IIR digital lter architectures which can perfor...
متن کاملEecient Implementations of Pipelined Cordic Based Iir Digital Filters Using Fast Orthonormal -rotations
CORDIC based IIR digital lters possess desirable properties for VLSI implementations such as regularity, local connection, low sensitivity to nite word-length implementation, and elimination of limit cycles. Recently, ne-grain pipelined CORDIC based IIR digital lter architectures which can perform the ltering operations at arbitrarily high sample rates at the cost of linear increase in hardware...
متن کاملA Cordic Argument Reduction Method with Regular Architecture and Unlimited Convergence Domain
One of the main problems of the CORDIC algorithm is the limited convergence domain, in which the functions can be calculated. Two different approaches can be employed to overcome this constraint: first, an argument reduction method and, second, an expansion of the CORDIC convergence domain. While the first approach requires significant processing overhead due to the need for divisions, the seco...
متن کاملA Low-Cost Floating Point Vectoring Algorithm Based on CORDIC∗
In this paper we develop a CORDIC-based floating-point vectoring algorithm which reduces significantly the amount of microrotation steps as compared to the conventional algorithm. The overhead required to accomplish this is minimized by the introduction of an angle selection function which considers only a few of the total amount of bits used to represent the vector being rotated. At the same t...
متن کاملComputation of Trigonometric Functions by the Systolic Implementation of the CORDIC Algorithm
Trigonometric functions are among the most useful functions in the digital signal processing applications. The design introduced in this paper computes the trigonometric functions by means of the systolic arrays. The method for computing these functions for an arbitrary angle, , is the CORDIC algorithm. A simple standard cell is used for the systolic array. Due to the fixed inputs, in some...
متن کامل