Full Text Available

Note: Clicking the button above will open the full text document at the original institutional repository in a new window.

AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism

Saved in:
Bibliographic Details
Published in:ArXiv cs.DC Recent Papers
Format: Online Article RSS Article
Published: 2026
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864130790313426946
collection WordPress RSS
FRELIP Feed Integration
container_title ArXiv cs.DC Recent Papers
description
discipline_display Engineering & Technology
discipline_facet Engineering & Technology
format Online Article
RSS Article
genre Journal Article
id rss_article:48781
institution FRELIP
journal_source_facet ArXiv cs.DC Recent Papers
publishDate 2026
publishDateSort 2026
record_format rss_article
spellingShingle AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
ArXiv cs.DC Recent Papers
Computer Science & IT
Engineering & Technology
sub_discipline_display Computer Science & IT
sub_discipline_facet Computer Science & IT
subject_display ArXiv cs.DC Recent Papers
Computer Science & IT
Engineering & Technology
ArXiv cs.DC Recent Papers
Computer Science & IT
Engineering & Technology
subject_facet ArXiv cs.DC Recent Papers
Computer Science & IT
Engineering & Technology
title AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_auth AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_full AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_fullStr AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_full_unstemmed AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_short AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
title_sort autosp: unlocking long-context llm training via compiler-based sequence parallelism
topic ArXiv cs.DC Recent Papers
Computer Science & IT
Engineering & Technology
url https://arxiv.org/abs/2604.27089v1