MelodyDeterminism: A Canonical Framework for Bitwise Reproducibility in LLM Inference
Year: 2025
Type: article
Abstract: Reproducibility is a key requirement in scientific research, regulatedindustries, and AI safety. Yet, large language models (LLMs) executed on modernaccelerators exhibit non-determinism due to parallel floating-point reductions,scheduling variability, and stochastic primitives. Current mitigations—such asdeterministic kernels or batch-invariant operators—address fragments of theproblem but lack a unifying framework. We propose MelodyDeterminism, analgorithmic approach where every numerical operator is treated as a melody:identified by a fingerprint (“intro”), mapped to a canonical rule (“score”), andexecuted in lockstep by GPU cores (“orchestra”). This ensures that identicalinputs always produce identical outputs, bitwise reproducible across runs,while minimizing overhead by applying determinism only where necessary.
Author: Vargiu, Simona
Language: English
Subfield: Hardware and Architecture
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7081572605
FWCI:
Cites:
Cited by:
Related to: