This work presents a technique for the resource optimization of input multiplexed ROM-based Finite State Machines. This technique exploits the don't care value of the inputs to reduce the memory size as well as multiplexer complexity. This technique has been applied to a publicly available FSM benchmarks and implemented in a low-cost FPGA. Results have been compared with tools supported ROM and standard logic cells implementations. In a significant number of test cases, the proposed technique is the best design alternative, both in resource requirements and speed.

Finite state machines with input multiplexing (FSMIMs) have been proposed in previous works as a technique for efficient mapping FSMs into ROM memory. In this paper, we propose a new architecture for implementing FSMIMs, called FSMIM with state-based input selection, whose goal is to achieve a further reduction in memory usage. This paper also describes in detail the algorithms for generating FSMIMs used by the tool FSMIM-Gen, which has been developed and made available on the Internet for free public use. A comparative study in terms of speed and area between FSMIM approaches and other field programmable gate array-based techniques is presented. The results show that the FSMIM approaches obtain huge reductions in the look-up table (LUT) usage by using a small number of embedded memory blocks. In addition, speed improvements over conventional LUT-based implementations have been obtained in many cases.

This paper presents a study of performance of RAM-based implementations in FPGAs of Finite State Machines (FSMs). The influence of the FSM characteristics on speed and area has been studied, taking into account the particular features of different FPGA families, like the size of LUTs, the size of memory blocks, the number of embedded multiplexer levels and the specific decoding logic for distributed RAM. Our study can be useful for efficiently implementing FPGA-based state machines.

A Binary-Tree-based Finite State Machine (BT-FSM) is a state machine with a 1-bit input signal whose state transition graph is a binary tree. BT-FSMs are useful in those application areas where searching in a binary tree is required, such as computer networks, compression, automatic control, or cryptography. This paper presents a new architecture for implementing BT-FSMs which is based on the model Finite Virtual State Machine (FVSM). The proposed architecture has been compared with the general FVSM and conventional approaches by using both synthetic test benches and very large BT-FSMs obtained from a real application. In synthetic test benches, the average speed improvement of the proposed architecture respect to the best results of the other approaches achieves 41% (there are some cases in which the speed is more than double). In the case of the real application, the average speed improvement achieves 155%.

Current Field Programmable Gate Arrays (FPGAs) provide fast routing links and special logic to perform carry operations; however, these resources can also be used to implement non-arithmetic circuits. In this paper, a new approach for mapping logic functions onto carry chains is presented. Unlike other approaches, the proposed technique can be applied to any logic function. The presented technique includes: (1) an architecture that is composed of blocks that implement AND and OR functions (called CANDs and CORs, respectively) by means of Look-Up-Tables (LUTs) and carry-chain resources; and (2) a mapping algorithm to reduce both the delay of the critical path and the number of used FPGA resources. The algorithm uses a heuristic to interconnect CORs and CANDs in order to reduce the delay. The problem of mapping the maxterms (or minterms) of a function to LUTs has been modelled as a Set Bin Packing (SBP) problem. Since SBP is NP-Hard, a greedy algorithm has been proposed, which is based on the First Fit Decreasing (FFD) heuristic. The results obtained have been compared with the conventional technique using both speed and area optimization. For this purpose, a large synthetic set of test cases has been generated. The proposed technique improves both the speed and area results for the vast majority of functions whose conventional implementation requires more than four logic levels. It is important to highlight that the improvement of one parameter (speed or area) is not achieved at the expense of the other.