GASLoC: New Decentralized Algorithm Improves Communication Efficiency in Large Language Model Pretraining
Researchers have introduced GASLoC, a decentralized pretraining algorithm designed to reduce communication bottlenecks when training large language models across distributed, heterogeneous compute environments. Unlike existing methods that rely on synchronous All-Reduce operations requiring identical model states across all workers, GASLoC uses gossip-based peer communication and supports local optimizer steps with adaptive optimizers. The work matters because communication overhead is an increasingly critical constraint as LLM training scales across data centers and lower-bandwidth network links.
As large language model pretraining increasingly spans multiple clusters and data centers with varying bandwidth, communication efficiency has become a significant bottleneck. Existing practical approaches reduce communication frequency but still depend on synchronous All-Reduce operations, which require all workers to maintain identical model states and can stall progress when hardware or network conditions are heterogeneous. GASLoC addresses this by generalizing communication acceleration to the 'outer optimizer' paradigm, enabling a gossip-based decentralized training framework compatible with adaptive optimizers and sparse randomized peer communication. Empirically, the authors report that GASLoC outperforms state-of-the-art decentralized algorithms in the single-step-per-communication setting across multiple network topologies. Notably, it is the first decentralized method in the LLM setting shown to match the performance of DiLoCo when using multiple local steps, and it demonstrates a clear advantage over DiLoCo in heterogeneous bandwidth scenarios. The paper spans 38 pages with 9 figures and was submitted to arXiv on June 9, 2026.
What's missing
The paper is a preprint and has not yet undergone peer review. Key open questions include how GASLoC scales to very large model sizes and cluster counts beyond those tested, and how sensitive performance is to the choice of peer communication sparsity and topology in real-world deployments.
What different sources said
- arXiv cs.LGCenter
Unifying Local Communications and Local Updates for LLM Pretraining
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.