Topics / Operating Systems
Weekly issue
Operating Systems Papers of the Week
The 20 most important new Operating Systems papers of the past week, ranked by potential impact. A new issue comes out every Monday, covering the previous Monday-to-Sunday week.
- 1Ripple: Deadline-Aware CPU-GPU Sharing for Embodied Inference
We present Ripple, a CPU-GPU co-scheduler that combines model-derived GPU memory volumes with CPU completion feedback to regulate interference.
- 2A Scoping Review of Agentic AI: Applications, Emerging Trends, Risks & Future Directions
A post hoc ACM Digital Library and Scopus conference-coverage sensitivity analysis showed that application-domain proportions were sensitive to database and document-type coverage, whereas the main recurring architectural signals remained qualitatively stable.
- 3UMLoad: Fast On-Demand Model Loading on Unified Memory Architecture
This paper presents UMLoad, a system-level model-loading accelerator that decouples a framework's logical allocations from their physical page backing.
- 4File-Semantic SSD for I/O-centric Architectures with NFS-over-NVMe Protocol
In this paper, we present File-Semantic SSD (FS-SSD), a new SSD architecture that provides a native file abstraction within SSDs, enabling direct file access by peripheral and remote devices without host intervention.
- 5QFabric: Guarded Cross-Domain Task Recovery and Its Limits on A Linux-RTOS Board
We present QFabric, a prototype that treats cross-domain placement as guarded recovery from a violated empirical soft real-time contract: a bounded task is compiled for both domains, monitored in windows with hysteresis, switched only after semantic and capacity admission at a quiescent epoch boundary, and committed or rolled back after probation.
Get papers like these the day they come out, not a week later.
Follow this category- 6RTP-LLM: Building An Evolvable and Reliable LLM Inference Engine at Alibaba Scale
To reconcile rapid evolution with production reliability, we present RTP-LLM, an open-source production inference engine developed and deployed at Alibaba.
- 7DIMWIT: Directly Invoking Models Without Intensive Tuning
We propose bypassing policy development and instead asking pretrained models to make resource-management decisions directly, given the current system state, desired outcome, and available choices.
- 8ToolSched: Resource-Efficient Tool Scheduling for Shared-Host LLM Agents
Thus, we present ToolSched, a Shell MCP Server for shared local hosts that serves multiple agents and schedules their operations.
- 9BitLLM: Efficient Low-Bit, Conversion-Hiding LLM Decoding on Mobile GPUs
In this paper, we identify that the performance gap stems primarily from frequent data type conversions between quantized and floating-point values, and dynamic and inefficient memory access patterns caused by the growing Key-Value (KV) cache during autoregressive decoding.
- 10Memory Fit: Architectural Support for Page Placement By Learning Models of Good Behavior
The paper describes soft trace matching (STM), a computationally and space efficient algorithm to compute memory fit.
- 11Blockchain-based PKI Infrastructure for Efficient Management of Medical Prescriptions
This paper introduces a decentralized cross-chain PKI architecture that integrates four complementary technologies: permissioned blockchains utilizing Practical Byzantine Fault Tolerance (PBFT) consensus, three functionally separated InterPlanetary File System (IPFS) nodes, a formally specified Lock & Mint cross-chain synchronization protocol deployed on a dedicated Ethereum relay chain, and entropy-verified secure QR codes.
- 12Interaction Between Access Cavity Design and NiTi Rotary Systems on Apical Debris Extrusion: An In Vitro Study
This study compared the TruNatomy (TRN) and RaceEvo (REV) files regarding the amount of apically extruded debris (AED) of mandibular molars with different access cavity designs: Traditional Endodontic Cavity (TEC) and Contracted Endodontic Cavity (CEC).
- 13When Benign Is Unverified: Label-Signal Mismatch and Abstention-First Triage in Graph-Based Memory Forensics
We report a validation study of a graph-based triage pipeline for Windows memory images: Volatility 3 output is compiled into typed heterogeneous graphs and scored by a temperature-scaled GINE classifier and two one-class encoders, fused with rule risk under an abstention gate.
- 14Beyond Source Compatibility: WebAssembly As A Stable ABI for ARINC 653 Applications
Our contribution is three-fold; first, we refine the ARINC 653 header file, eliminating remaining ambiguities.
- 15EBPF-Driven Adaptive Lifetime-Aware Data Placement for Shared FDP SSDs
We present eFDP, an eBPF-based system that automatically learns data lifetimes and manages FDP placement for unmodified applications on a stock Linux kernel.
- 16High-Throughput Private Information Retrieval with On-CPU Accelerators
In this paper, we present AltoPIR, the first PIR system that leverages on-CPU matrix accelerators such as Intel AMX.
- 17Passkeys and Fido2: Security, Usability and Deployment of Passwordless Authentication
We describe how a signature bound to the web origin defeats phishing, then set out what formal analyses have proved and under which assumptions.
- 18Sistemas Híbridos De Rastreamento De Contêineres
The objective is to propose a conceptual and operational model for transforming containers into smart assets, capable of self-tracking on land and deep sea, and predictive integration with customs yards to minimize unproductive movements.
- 19DDoS Attack Detection on Portmap Using A Bidirectional Recurrent Neural Network Method
This study uses the CSE-CIC-IDS 2019 dataset, especially the DDoS attack dataset on portmap.
- 20Containerized Runtime Berbasis Docker Untuk Lingkungan Praktikum Pemrograman Yang Seragam Dan Terisolasi
This study implements a Docker-based containerized runtime in a web-based interactive practicum platform using the Rapid Application Development (RAD) method, covering requirements planning, system design, development, implementation, and testing.