@inproceedings{nie-etal-2026-sfmp,
title = "{SFMP}: Fine-Grained, Hardware-Friendly and Solver-Free Mixed-Precision Quantization for Large Language Models",
author = "Nie, Xin and
Zhang, Haicheng and
Dong, Liang and
Feng, Beining and
Weng, Jinhong and
Sun, Guiling",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.113/",
pages = "2233--2259",
ISBN = "979-8-89176-444-6",
abstract = "Mixed-precision quantization is a promising approach for compressing large language models under tight memory budgets. However, existing mixed-precision methods typically suffer from one of two limitations: they either rely on expensive discrete optimization to determine precision allocation, or introduce hardware inefficiencies due to irregular memory layouts. We propose SFMP, a solver-free and hardware-friendly mixed-precision quantization framework for large language models. SFMP defines a theoretically simple objective for mixed-precision quantization and is built upon three novel ideas: 1) Block-wise mixed-precision, enabling fine-grained precision within weight matrices while remaining hardware-friendly; 2) Row-column weight reordering, which improves structural alignment via row and column reordering, incurring only a small activation reordering overhead during inference; 3) Unified GEMM kernel, which supports block-wise mixed-precision GEMM at arbitrary average bit-widths. Extensive experiments demonstrate that SFMP outperforms state-of-the-art layer-wise mixed-precision methods under the same memory constraints, while significantly reducing quantization cost and improving inference efficiency. Code is available at https://github.com/Nkniexin/SFMP."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nie-etal-2026-sfmp">
<titleInfo>
<title>SFMP: Fine-Grained, Hardware-Friendly and Solver-Free Mixed-Precision Quantization for Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xin</namePart>
<namePart type="family">Nie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haicheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liang</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Beining</namePart>
<namePart type="family">Feng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinhong</namePart>
<namePart type="family">Weng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guiling</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>Mixed-precision quantization is a promising approach for compressing large language models under tight memory budgets. However, existing mixed-precision methods typically suffer from one of two limitations: they either rely on expensive discrete optimization to determine precision allocation, or introduce hardware inefficiencies due to irregular memory layouts. We propose SFMP, a solver-free and hardware-friendly mixed-precision quantization framework for large language models. SFMP defines a theoretically simple objective for mixed-precision quantization and is built upon three novel ideas: 1) Block-wise mixed-precision, enabling fine-grained precision within weight matrices while remaining hardware-friendly; 2) Row-column weight reordering, which improves structural alignment via row and column reordering, incurring only a small activation reordering overhead during inference; 3) Unified GEMM kernel, which supports block-wise mixed-precision GEMM at arbitrary average bit-widths. Extensive experiments demonstrate that SFMP outperforms state-of-the-art layer-wise mixed-precision methods under the same memory constraints, while significantly reducing quantization cost and improving inference efficiency. Code is available at https://github.com/Nkniexin/SFMP.</abstract>
<identifier type="citekey">nie-etal-2026-sfmp</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.113/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>2233</start>
<end>2259</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SFMP: Fine-Grained, Hardware-Friendly and Solver-Free Mixed-Precision Quantization for Large Language Models
%A Nie, Xin
%A Zhang, Haicheng
%A Dong, Liang
%A Feng, Beining
%A Weng, Jinhong
%A Sun, Guiling
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F nie-etal-2026-sfmp
%X Mixed-precision quantization is a promising approach for compressing large language models under tight memory budgets. However, existing mixed-precision methods typically suffer from one of two limitations: they either rely on expensive discrete optimization to determine precision allocation, or introduce hardware inefficiencies due to irregular memory layouts. We propose SFMP, a solver-free and hardware-friendly mixed-precision quantization framework for large language models. SFMP defines a theoretically simple objective for mixed-precision quantization and is built upon three novel ideas: 1) Block-wise mixed-precision, enabling fine-grained precision within weight matrices while remaining hardware-friendly; 2) Row-column weight reordering, which improves structural alignment via row and column reordering, incurring only a small activation reordering overhead during inference; 3) Unified GEMM kernel, which supports block-wise mixed-precision GEMM at arbitrary average bit-widths. Extensive experiments demonstrate that SFMP outperforms state-of-the-art layer-wise mixed-precision methods under the same memory constraints, while significantly reducing quantization cost and improving inference efficiency. Code is available at https://github.com/Nkniexin/SFMP.
%U https://aclanthology.org/2026.findings-aacl.113/
%P 2233-2259
Markdown (Informal)
[SFMP: Fine-Grained, Hardware-Friendly and Solver-Free Mixed-Precision Quantization for Large Language Models](https://aclanthology.org/2026.findings-aacl.113/) (Nie et al., Findings 2026)
ACL