arXiv Machine Learning October 16, 2026 By Tingfeng Lan, Yusen Wu, Bin Ma, Zhaoyuan Su, Rui Yang, Tekin Bicer, Masahiro Tanaka, Olatunji Ruwase, Dong Li, Yue Cheng

ZenFlow: Enabling Stall-Free Offloading for LLM Training

Read the original on arXiv Machine Learning →

ZenFlow is an offloaded training framework for large language models that reduces GPU stalls by prioritizing important gradients. It updates critical gradients in-place on the GPU while offloading less important ones to the CPU, overlapping CPU work with GPU computation. The approach leverages spatio‑temporal locality of important gradients to keep identification lightweight, achieving up to 5× speedup, 2× less PCIe traffic, 85% fewer GPU stalls, and lower cloud GPU costs without sacrificing accuracy.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv Machine Learning

Aug 24

DAOP: Data-Aware Offloading and Predictive Pre-Calculation for Efficient MoE Inference

arXiv:2501.10375v3 Announce Type: replace-cross Abstract: Mixture-of-Experts (MoE) models, though highly effective for various machine learning tasks, face significant deployment challenges on memory...

By Yujie Zhang, Shivam Aggarwal, Tulika Mitra

arXiv AI

5d ago

First-Order LoRA Training with Phone NPUs

The paper introduces a system that enables first‑order LoRA training on phone NPUs by running all frozen‑weight multiplications on the NPU while keeping adapters and the optimizer on the CPU. It reuses weight tiles for input‑gradient computation, selects precision based on gradient sensitivity, and fuses remaining CPU operations to overlap with NPU execution. The approach is evaluated on Qualcomm Snapdragon devices with Qwen and Llama LLMs, demonstrating background training for nightly personalization and benchmark workloads.

By Xu Zheng, Wei Cheng, Zhuomin Chen, Dongkuan Xu, Mo Sha, Jingchao Ni, Dongsheng Luo

llms fine-tuning benchmarks

arXiv AI

Jun 24

CrossPool: Efficient Multi-LLM Serving for Cold MoE Models through KV-Cache and Weight Disaggregation

arXiv:2606. 24506v1 Announce Type: cross Abstract: Emerging LLM services increasingly host many sparse MoE models, yet most models receive sparse requests and remain cold.

By Zhuoren Ye, Tianyu Wo, Dinghao Xue, Mingming Zhang, Yuchen Teng, Chunming Hu, Renyu Yang

llms benchmarks

arXiv AI

Jul 7

Data Driven Optimization of GPU efficiency for Distributed LLM-Adapter Serving

arXiv:2602. 24044v2 Announce Type: replace-cross Abstract: Large Language Model (LLM) adapters enable low-cost model specialization, but introduce complex caching and scheduling challenges in distributed serving systems where hundreds of adapters must be hosted concurrently.

By Ferran Agullo, Joan Oliveras, Chen Wang, Alberto Gutierrez-Torre, Olivier Tardieu, Alaa Youssef, Jordi Torres, Josep Ll. Berral

llms efficiency benchmarks

arXiv AI

Jul 1

An Efficient Heterogeneous Co-Design for Fine-Tuning on a Single GPU

arXiv:2603. 16428v2 Announce Type: replace-cross Abstract: Fine-tuning Large Language Models (LLMs) has become essential for domain adaptation, but its memory-intensive property exceeds the capabilities of most GPUs.

By Ruijia Yang, Zeyi Wen

llms fine-tuning efficiency

arXiv AI

Jul 21

KernelBench-Verified: Do LLM-Generated Kernels Actually Beat PyTorch?

arXiv:2607. 16241v1 Announce Type: cross Abstract: Recent large language models (LLMs) can generate custom CUDA kernels that appear to outperform PyTorch on benchmarks such as KernelBench.

By Yunxiang Zhang (Xiangjun), Ping Yu (Xiangjun), Jianyu Wang (Xiangjun), Max (Xiangjun), Fan, Julian Reed, Azalia Mirhoseini, Will Su

llms benchmarks

One email a morning, machine-written

Related stories

DAOP: Data-Aware Offloading and Predictive Pre-Calculation for Efficient MoE Inference

First-Order LoRA Training with Phone NPUs

CrossPool: Efficient Multi-LLM Serving for Cold MoE Models through KV-Cache and Weight Disaggregation

Data Driven Optimization of GPU efficiency for Distributed LLM-Adapter Serving

An Efficient Heterogeneous Co-Design for Fine-Tuning on a Single GPU

KernelBench-Verified: Do LLM-Generated Kernels Actually Beat PyTorch?