@inproceedings{he-etal-2026-much,
title = "How Much Parallelism Is ``Free''? A Principle of Near-Free Parallelism for Parallel Decoding",
author = "He, Minghua and
Zhang, Lingzhe and
Liu, Yuan and
Xiao, Zhou and
Liu, Aiwei",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2046/",
pages = "43520--43578",
ISBN = "979-8-89176-462-0",
abstract = "Parallel decoding improves generation efficiency by processing multiple decode positions within a single decode forward, but reported speedups conflate algorithmic token utilization with the system cost of executing multiple positions. We isolate the system side by introducing Near-Free Parallelism (NFP), the maximum number of positions executable at near-free latency. Analyzing Dense FFNs, MoE FFNs, and Attention against an idle-compute baseline, we find that NFP is shaped not by memory-bound resource slack alone, but also by implementation-induced kernel-granularity slack. Based on these mechanisms, we establish a Near-Free Parallelism principle that predicts the NFP boundary from hardware balance and implementation granularity. Validation on representative Dense and MoE models{---}spanning both diffusion and autoregressive decoding{---}shows that the principle accurately predicts practical NFP boundaries, revealing that the standard idle-compute intuition over-predicts them by 2.4{--}23x in over half of the evaluated deployment settings{---}offering a system-side budget for parallelism selection and model-system co-design."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="he-etal-2026-much">
<titleInfo>
<title>How Much Parallelism Is “Free”? A Principle of Near-Free Parallelism for Parallel Decoding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Minghua</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingzhe</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuan</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhou</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aiwei</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Parallel decoding improves generation efficiency by processing multiple decode positions within a single decode forward, but reported speedups conflate algorithmic token utilization with the system cost of executing multiple positions. We isolate the system side by introducing Near-Free Parallelism (NFP), the maximum number of positions executable at near-free latency. Analyzing Dense FFNs, MoE FFNs, and Attention against an idle-compute baseline, we find that NFP is shaped not by memory-bound resource slack alone, but also by implementation-induced kernel-granularity slack. Based on these mechanisms, we establish a Near-Free Parallelism principle that predicts the NFP boundary from hardware balance and implementation granularity. Validation on representative Dense and MoE models—spanning both diffusion and autoregressive decoding—shows that the principle accurately predicts practical NFP boundaries, revealing that the standard idle-compute intuition over-predicts them by 2.4–23x in over half of the evaluated deployment settings—offering a system-side budget for parallelism selection and model-system co-design.</abstract>
<identifier type="citekey">he-etal-2026-much</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2046/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>43520</start>
<end>43578</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T How Much Parallelism Is “Free”? A Principle of Near-Free Parallelism for Parallel Decoding
%A He, Minghua
%A Zhang, Lingzhe
%A Liu, Yuan
%A Xiao, Zhou
%A Liu, Aiwei
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F he-etal-2026-much
%X Parallel decoding improves generation efficiency by processing multiple decode positions within a single decode forward, but reported speedups conflate algorithmic token utilization with the system cost of executing multiple positions. We isolate the system side by introducing Near-Free Parallelism (NFP), the maximum number of positions executable at near-free latency. Analyzing Dense FFNs, MoE FFNs, and Attention against an idle-compute baseline, we find that NFP is shaped not by memory-bound resource slack alone, but also by implementation-induced kernel-granularity slack. Based on these mechanisms, we establish a Near-Free Parallelism principle that predicts the NFP boundary from hardware balance and implementation granularity. Validation on representative Dense and MoE models—spanning both diffusion and autoregressive decoding—shows that the principle accurately predicts practical NFP boundaries, revealing that the standard idle-compute intuition over-predicts them by 2.4–23x in over half of the evaluated deployment settings—offering a system-side budget for parallelism selection and model-system co-design.
%U https://aclanthology.org/2026.findings-emnlp.2046/
%P 43520-43578
Markdown (Informal)
[How Much Parallelism Is "Free"? A Principle of Near-Free Parallelism for Parallel Decoding](https://aclanthology.org/2026.findings-emnlp.2046/) (He et al., Findings 2026)
ACL