F-BFQ: Flexible Block Floating-Point Quantization Accelerator for LLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Haris, Jude, Cano, José
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912648739684352
author Haris, Jude
Cano, José
author_facet Haris, Jude
Cano, José
contents Large Language Models (LLMs) have become increasingly prominent for daily tasks, from improving sound-totext translation to generating additional frames for the latest video games. With the help of LLM inference frameworks, such as llama.cpp, which support optimizations such as KV-caching and quantization, it is now easier than ever to deploy LLMs on edge devices. Quantization is fundamental to enable LLMs on resource-constrained edge devices, and llama.cpp utilizes block floating point (BFP) quantization to drastically reduce the bit width of weights and input tensors, the memory footprint, and the computational power required to run LLMs. LLMs are typically quantized with mixed BFP quantization across the model layers to reduce the loss of model accuracy due to quantization. Therefore, to efficiently accelerate across the layers of BFP-quantized LLMs, specialized accelerators need to support different BFP variants without reconfiguration. To address this issue, we propose a Flexible Block FloatingPoint Quantization (F-BFQ) accelerator, which can dynamically switch between two BFP quantization variants and perform matrix multiplication (MatMul) operations. Our initial F-BFQ accelerator design, deployed on the AMD Kria board, reduces inference time by 1.4x on average over the Arm NEON-based CPU execution across three BFP quantized LLMs while achieving 5.2 tokens per second (~3.9 words per second).
format Preprint
id arxiv_https___arxiv_org_abs_2510_13401
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle F-BFQ: Flexible Block Floating-Point Quantization Accelerator for LLMs
Haris, Jude
Cano, José
Hardware Architecture
Distributed, Parallel, and Cluster Computing
Machine Learning
Large Language Models (LLMs) have become increasingly prominent for daily tasks, from improving sound-totext translation to generating additional frames for the latest video games. With the help of LLM inference frameworks, such as llama.cpp, which support optimizations such as KV-caching and quantization, it is now easier than ever to deploy LLMs on edge devices. Quantization is fundamental to enable LLMs on resource-constrained edge devices, and llama.cpp utilizes block floating point (BFP) quantization to drastically reduce the bit width of weights and input tensors, the memory footprint, and the computational power required to run LLMs. LLMs are typically quantized with mixed BFP quantization across the model layers to reduce the loss of model accuracy due to quantization. Therefore, to efficiently accelerate across the layers of BFP-quantized LLMs, specialized accelerators need to support different BFP variants without reconfiguration. To address this issue, we propose a Flexible Block FloatingPoint Quantization (F-BFQ) accelerator, which can dynamically switch between two BFP quantization variants and perform matrix multiplication (MatMul) operations. Our initial F-BFQ accelerator design, deployed on the AMD Kria board, reduces inference time by 1.4x on average over the Arm NEON-based CPU execution across three BFP quantized LLMs while achieving 5.2 tokens per second (~3.9 words per second).
title F-BFQ: Flexible Block Floating-Point Quantization Accelerator for LLMs
topic Hardware Architecture
Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2510.13401