Revealing Floating-Point Accumulation Orders in Software/Hardware Implementations

TL;DR

This paper introduces FPRev, a diagnostic tool that non-intrusively reveals the undocumented accumulation orders in software and hardware implementations, aiding reproducibility and verification across diverse computational platforms.

Contribution

The paper presents FPRev, a novel testing-based tool that efficiently uncovers accumulation orders in various implementations without modification.

Findings

FPRev can identify accumulation orders in popular libraries like NumPy and PyTorch.

FPRev works effectively on CPUs, GPUs, and specialized accelerators.

The tool demonstrates lower time complexity than basic solutions.

Abstract

Accumulation-based operations, such as summation and matrix multiplication, are fundamental to numerous computational domains. However, their accumulation orders are often undocumented in existing software and hardware implementations, making it difficult for developers to ensure consistent results across systems. To address this issue, we introduce FPRev, a diagnostic tool designed to reveal the accumulation order in the software and hardware implementations through numerical testing. With FPRev, developers can identify and compare accumulation orders, enabling developers to create reproducible software and verify implementation equivalence. FPRev is a testing-based tool that non-intrusively reveals the accumulation order by analyzing the outputs of the tested implementation for distinct specially designed inputs. Employing FPRev, we showcase the accumulation orders of popular libraries (such as NumPy and PyTorch) on CPUs and GPUs (including GPUs with specialized matrix accelerators such as Tensor Cores). We also validate the efficiency of FPRev through extensive experiments. FPRev exhibits a lower time complexity compared to the basic solution. FPRev is open-sourced at https://github.com/peichenxie/FPRev.