CICERO: A Domain-Specific Architecture for Efficient Regular Expression Matching

Regular Expression (RE) matching is a computational kernel used in several applications. Since RE complexity and data volumes are steadily increasing, hardware acceleration is gaining attention also for this problem. Existing approaches have limited flexibility as they require a different implementation for each RE. On the other hand, it is complex to map efficient RE representations like non-deterministic finite-state automata onto software-programmable engines or parallel architectures. In this work, we present CICERO , an end-to-end framework composed of a domain-specific architecture and a companion compilation framework for RE matching. Our solution is suitable for many applications, such as genomics/proteomics and natural language processing. CICERO aims at exploiting the intrinsic parallelism of non-deterministic representations of the REs. CICERO can trade-off accelerators’ efficiency and processors’ flexibility thanks to its programmable architecture and the compilation framework. We implemented CICERO prototypes on embedded FPGA achieving up to 28.6× and 20.8× more energy efficiency than embedded and mainstream processors, respectively. Since it is a programmable architecture, it can be implemented as a custom ASIC that is orders of magnitude more energy-efficient than mainstream processors.

Resonance for analog recurrent neural network

There is a strong interest in using physical waves for artificial neural computing because of their unique advantages in fast speed and intrinsic parallelism. Resonance, as a ubiquitous feature across many wave systems, is a natural candidate for analog computing in temporal signals. We demonstrate that resonance can be used to construct stable and scalable recurrent neural networks. By including resonators with different lifetimes, the computing system develops both short-term and long-term memory simultaneously.

An Efficient Parallel Approximate Algorithm for Solving Time Fractional Reaction-Diffusion Equations

In this paper, we construct pure alternative segment explicit-implicit (PASE-I) and implicit-explicit (PASI-E) difference algorithms for time fractional reaction-diffusion equations (FRDEs). They are a kind of difference schemes with intrinsic parallelism and based on classical explicit scheme and classical implicit scheme combined with alternating segment technology. The existence and uniqueness analysis of solutions of the parallel difference schemes are given. Both the theoretical proof and the numerical experiment show that PASE-I and PASI-E schemes are unconditionally stable and convergent with second-order spatial accuracy and 2−α order time accuracy. Compared with implicit scheme and E-I (I-E) scheme, the computational efficiency of PASE-I and PASI-E schemes is greatly improved. PASE-I and PASI-E schemes have obvious parallel computing properties, which shows that the difference schemes with intrinsic parallelism in this paper are feasible to solve the time FRDEs.

A New Class of Difference Methods with Intrinsic Parallelism for Burgers–Fisher Equation

This paper proposes a new class of difference methods with intrinsic parallelism for solving the Burgers–Fisher equation. A new class of parallel difference schemes of pure alternating segment explicit-implicit (PASE-I) and pure alternating segment implicit-explicit (PASI-E) are constructed by taking simple classical explicit and implicit schemes, combined with the alternating segment technique. The existence, uniqueness, linear absolute stability, and convergence for the solutions of PASE-I and PASI-E schemes are well illustrated. Both theoretical analysis and numerical experiments show that PASE-I and PASI-E schemes are linearly absolute stable, with 2-order time accuracy and 2-order spatial accuracy. Compared with the implicit scheme and the Crank–Nicolson (C-N) scheme, the computational efficiency of the PASE-I (PASI-E) scheme is greatly improved. The PASE-I and PASI-E schemes have obvious parallel computing properties, which show that the difference methods with intrinsic parallelism in this paper are feasible to solve the Burgers–Fisher equation.

The Splitting Crank–Nicolson Scheme with Intrinsic Parallelism for Solving Parabolic Equations

In this paper, a splitting Crank–Nicolson (SC-N) scheme with intrinsic parallelism is proposed for parabolic equations. The new algorithm splits the Crank–Nicolson scheme into two domain decomposition methods, each one is applied to compute the values at (n + 1)th time level by use of known numerical solutions at n-th time level, respectively. Then, the average of the above two values is chosen to be the numerical solutions at (n + 1)th time level. The new algorithm obtains accuracy of the Crank–Nicolson scheme while maintaining parallelism and unconditional stability. This algorithm can be extended to solve two-dimensional parabolic equations by alternating direction implicit (ADI) technique. Numerical experiments illustrate the accuracy and efficiency of the new algorithm.

A New Alternating Segment Crank-Nicolson Scheme for the Fourth-Order Parabolic Equation

A group of asymmetric difference schemes to approach the fourth-order parabolic equation is given. According to these schemes and the Crank-Nicolson scheme, an alternating segment Crank-Nicolson scheme with intrinsic parallelism is constructed. The truncation errors and the stability are discussed. Numerical simulations show that this new scheme has unconditional stability and high accuracy and convergency, and it is in preference to the implicit scheme method.