paper-with-me

Papers

Throughput Optimization as a Strategic Lever in Large-Scale AI Systems: Evidence from Dataloader and Memory Profiling Innovations

2026-03-27 · Mayank Jha arxiv

The development of large-scale foundation models, particularly Large Language Models (LLMs), is constrained by significant computational and memory bottlenecks. These challenges elevate throughput optimization from a mere engineering task to a critical strategic lever, directly influencing training time, operational cost, and the feasible scale of next-generation models. This paper synthesizes evidence from recent academic and industry innovations to analyze key advancements in training efficiency. We examine architectural solutions to dataloader bottlenecks, such as the OVERLORD framework, which has demonstrated a 4.5% improvement in end-to-end training throughput. We investigate memory optimization techniques designed to overcome the GPU memory wall, including CPU offloading strategies like DeepSpeed's ZeRO-Offload, which enable the training of models far exceeding single-accelerator capacity. Furthermore, we explore the growing importance of compiler-centric optimizations, exemplified by Triton-distributed, which enables the joint optimization of computation, memory, and communication for substantial performance gains. The analysis is contextualized by advanced profiling tools and hardware characterization studies that identify and mitigate previously overlooked overheads like Dynamic Voltage and Frequency Scaling (DVFS). Findings indicate that a holistic, system-level approach, integrating innovations across data pipelines, memory management, network fabrics, and compiler technologies, is essential for accelerating AI development, managing costs, and pushing the boundaries of model scale.

📄 PDF Abstract BibTeX arXiv:2603.26823

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

HaShiFlex: A High-Throughput Hardened Shifter DNN Accelerator with Fine-Tuning Flexibility

2025-12-14 · Jonathan Herbst, Michael Pellauer, Sherief Reda arxiv

We introduce a high-throughput neural network accelerator that embeds most network layers directly in hardware, minimizing data transfer and memory usage while preserving a degree of flexibility via a small neural proces…

ParaFold: Paralleling AlphaFold for Large-Scale Predictions

2021-11-11 · Bozitao Zhong, Xiaoming Su, Minhua Wen, Sichen Zuo 외

AlphaFold predicts protein structures from the amino acid sequence at or near experimental resolution, solving the 50-year-old protein folding challenge, leading to progress by transforming large-scale genomics data into…

CPUGPUProtein Folding

Breaking the Gradient Barrier: Unveiling Large Language Models for Strategic Classification

2025-11-10 · Xinpeng Lv, Yunxin Mao, Haoxuan Li, Ke Liang 외 arxiv

Strategic classification~(SC) explores how individuals or entities modify their features strategically to achieve favorable classification outcomes. However, existing SC methods, which are largely based on linear models …

JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency

2026-04-03 · Aichen Cai, Anmeng Zhang, Anyu Li, Bo Zhang 외 arxiv

We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performance and token efficiency in the sub-50B parameter regime. JoyAI-LLM Flash is pr…

Reinforcement Learning

An Efficient Continuous-Time MILP for Integrated Aircraft Hangar Scheduling and Layout

2025-08-04 · Shayan Farhang Pazhooh, Hossein Shams Shemirani arxiv

Efficient management of aircraft MRO hangars requires the integration of spatial layout with time-continuous scheduling to minimize operational costs. We propose a continuous-time mixed-integer linear program that jointl…