ПРЕПРИНТ

Эта статья является препринтом и не была отрецензирована.
О результатах, изложенных в препринтах, не следует сообщать в СМИ как о проверенной информации.
QUEFI: A Quaternion Firmware Interface with Integrated Neural Core (QNCORE) for Hardware State Representation
2026-03-19

We present QUEFI and QNCORE — a quaternion firmware interface with an integrated neural core for hardware state representation. QUEFI defines a quaternion-valued device abstraction layer (QBus) where 28 hardware drivers expose status, command, and compute interfaces as quaternion-valued functions. QNCORE is an SSE-accelerated in-kernel neural inference engine. The system is implemented as a 96,689 LOC bare-metal x86 kernel. Key results: 18–34% speed advantage at 4× parameter compression, 94.0% fault detection accuracy at 17× fewer parameters, 65× higher class separability via Hamilton aggregation, and 3.6× latency reduction from in-kernel placement. Files

Ссылка для цитирования:

Dzigirej A. 2026. QUEFI: A Quaternion Firmware Interface with Integrated Neural Core (QNCORE) for Hardware State Representation. PREPRINTS.RU. https://doi.org/10.24108/preprints-3114726

Список литературы