Abstract:Stochastic computing is a paradigm in which logical operations are performed on randomly generated bit streams. Complex arithmetic operations can be executed by simple logic circuits, resulting in a much smaller area footprint compared to conventional binary counterparts. However, the random or pseudorandom sources required for generating the bit streams are costly in terms of area and offset the advantages. Additionally, due to the inherent randomness, the computation lacks precision, limiting the applicability of this paradigm. Importantly, achieving reasonable accuracy in stochastic computing involves high latency. Recently, deterministic approaches to stochastic computing have been proposed, demonstrating that randomness is not a requirement. By structuring the computation deterministically, exact results can be obtained, and the latency greatly reduced. The bit stream generated adheres to a "unary" encoding, retaining the non-positional nature of the bits while discarding the random bit generation of traditional stochastic computing. This deterministic approach overcomes many drawbacks of stochastic computing, although the latency increases quadratically with each level of logic, becoming unmanageable beyond a few levels. In this paper, we present a method for approximating the results of the deterministic method while maintaining low latency at each level. This improvement comes at the cost of additional logic, but we demonstrate that the increase in area scales with the square root of n, where n represents the equivalent number of binary bits of precision. Our new approach is general, efficient, composable, and applicable to all arithmetic operations performed with stochastic logic. We show that this approach outperforms other stochastic designs for matrix multiplication (dot-product), which is an integral step in nearly all machine learning algorithms.

Weighted-Adder-Based Polynomial Computation Using Correlated Unipolar Stochastic Bitstreams

A Universal Approximation Method and Optimized Hardware Architectures for Arithmetic Functions Based on Stochastic Computing

Stochastic Synthesis for Stochastic Computing

A Reconfigurable Multiplier for Signed Multiplications with Asymmetric Bit-Widths.

A Scalable Approach to Performing Multiplication and Matrix Dot-Products in Unary

A Parallel Bitstream Generator for Stochastic Computing

From Multipliers to Integrators: A Survey of Stochastic Computing Primitives

Efficient Non-Linear Adder for Stochastic Computing with Approximate Spatial-Temporal Sorting Network

Accurate and Energy-Efficient Implementation of Non-Linear Adder in Parallel Stochastic Computing Using Sorting Network

From Stochastic to Bit Stream Computing: Accurate Implementation of Arithmetic Circuits and Applications in Neural Networks

Low-Cost and Highly-Efficient Bit-Stream Generator for Stochastic Computing Division

Design of a Stochastic Computing Architecture for the Phansalkar Algorithm

Efficient Parallel Stochastic Computing Multiply-Accumulate (MAC) Technique Using Pseudo-Sobol Bit-Streams

Weighted p -Bits for FPGA Implementation of Probabilistic Circuits

An Accurate and Time-Efficient Subtractor by Cross Format Coding in Stochastic Computing

CeMux: Maximizing the Accuracy of Stochastic Mux Adders and an Application to Filter Design

Parallel Hybrid Stochastic-Binary-Based Neural Network Accelerators

Stochastic Computing Hardware Design and Optimization for Convolutional Neutral Networks

Simultaneous Area and Latency Optimization for Stochastic Circuits by D Flip-Flop Insertion

CMOS plus stochastic nanomagnets enabling heterogeneous computers for probabilistic inference and learning

Stochastic Computing for Sub-0.2V Applications-Technology Evaluation and Design Optimization