Customizable Precision of Floating-Point Arithmetic with Bitslice Vector Types

TL;DR

This paper introduces a flexible vector computing method using bitslice format for custom-precision floating point data, enabling efficient software-based arithmetic with adjustable precision levels.

Contribution

It presents a novel bitslice vector representation and bitwise instruction set for customizable floating point precision in software, optimizing resource use.

Findings

Efficient for low-precision floating point vectors

Supports arbitrary bit precision

Demonstrates practical performance benefits

Abstract

Customizing the precision of data can provide attractive trade-offs between accuracy and hardware resources. We propose a novel form of vector computing aimed at arrays of custom-precision floating point data. We represent these vectors in bitslice format. Bitwise instructions are used to implement arithmetic circuits in software that operate on customized bit-precision. Experiments show that this approach can be efficient for vectors of low-precision custom floating point types, while providing arbitrary bit precision.