← 研踪 PaperTrailarXiv:2611.04253 · v1
UMLoad: Fast On-Demand Model Loading on Unified Memory Architecture
首次提交 2026-11-03 11:44
版本更新 2026-11-03 11:44
雷达观测 2026-11-04 23:40
摘要
On-demand model loading dominates response latency for interactive on-device AI, yet unified memory architectures (UMAs), now common on edge devices, should make it cheap by placing weights in memory the GPU already sees. However, frameworks still use the loading flow built for discrete GPUs, leaving significant inefficiencies. This paper presents UMLoad, a system-level model-loading accelerator that decouples a framework's logical allocations from their physical page backing. UMLoad uses on-demand reconstruction to eliminate the weight copy while preserving copy independence, and GPU-assisted page zeroing to accelerate runtime allocation while preserving its security guarantee. Implemented on Intel, AMD, and NVIDIA UMA platforms and evaluated with llama.cpp, PyTorch, and vLLM across six models, UMLoad reduces average loading time by 36.6%-71.9% over vanilla and 19.9%-62.7% over pipelined loaders, and cuts mean LLM-routing time to first token (TTFT) by 51.3%-59.3% and model-backed tool completion latency by 10.2%-31.6%.
作者备注：17 pages, 12 figures, 2 tables
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-03 11:44 · 最新观测版本
UMLoad: Fast On-Demand Model Loading on Unified Memory Architecture
On-demand model loading dominates response latency for interactive on-device AI, yet unified memory architectures (UMAs), now common on edge devices, should make it cheap by placing weights in memory the GPU already sees. However, frameworks still use the loading flow built for discrete GPUs, leaving significant inefficiencies. This paper presents UMLoad, a system-level model-loading accelerator that decouples a framework's logical allocations from their physical page backing. UMLoad uses on-demand reconstruction to eliminate the weight copy while preserving copy independence, and GPU-assisted page zeroing to accelerate runtime allocation while preserving its security guarantee. Implemented on Intel, AMD, and NVIDIA UMA platforms and evaluated with llama.cpp, PyTorch, and vLLM across six models, UMLoad reduces average loading time by 36.6%-71.9% over vanilla and 19.9%-62.7% over pipelined loaders, and cuts mean LLM-routing time to first token (TTFT) by 51.3%-59.3% and model-backed tool completion latency by 10.2%-31.6%.
查看此版本解读 →查看此版本原文 ↗