This paper presents ScaleRAN, a runtime for elastic GPU resource management in AI-RAN, enabling dynamic partitioning of GPU resources between RAN and AI workloads.
This paper presents a novel runtime system for elastic GPU resource management in AI-RAN, which is different from existing GPU-sharing mechanisms that rely on coarse-grained hardware partitioning or static resource allocation.
Before reading this…
Applications
To understand this paper, make sure you know these concepts first:
AI-RAN is driving the convergence of radio access network (RAN) processing and artificial intelligence (AI) workloads on GPU-enabled computing platforms, ranging from GPU-accelerated base stations to cloud-native infrastructures. While GPU acceleration improves the efficiency of virtualized RAN functions, the computational demand of latency-critical baseband processing varies with traffic load, creating transient spare GPU capacity that can be reclaimed by co-located AI applications. Existing GPU-sharing mechanisms rely on coarse-grained hardware partitioning or static resource allocation, limiting their ability to adapt to slot-level workload fluctuations while preserving strict timing requirements. This paper presents ScaleRAN, a node-local runtime for elastic GPU resource management in AI-RAN. ScaleRAN leverages NVIDIA Green Contexts to dynamically partition Streaming Multiprocessors (SMs) between RAN and AI workloads through a ladder of pre-created GPU partitions, enabling low-overhead resource reallocation at RAN slot granularity. The proposed runtime naturally supports Open-RAN (O-RAN) O-Cloud deployments, one representative AI-RAN deployment scenario. To drive these reallocations, we propose a reactive controller inspired by Kubernetes autoscaling and a proactive controller based on quantile regression and conformal prediction. We also develop LDPCSim, a GPU-accelerated workload generator built on NVIDIA Aerial for reproducible evaluation. Results show that predictive fine-grained GPU partitioning satisfies slot-level RAN deadlines while increasing the GPU resources available to co-located AI workloads, demonstrating the practicality of elastic GPU sharing for future AI-RAN platforms.
Beaver: Elastic GPU Sharing between ML and Latency-Critical vRAN Workloads
This paper presents Beaver, a GPU sharing system that jointly manages compute an…
Cross-Layer Intrusion Detection in 5G O-RAN: Gains and Limits of Fusing Radio Telemetry with Network…
This paper evaluates the effectiveness of fusing radio telemetry and CU-side net…
gKSM: GPU Memory Deduplication at Driver Layer for Serverless LLM Inference
gKSM is a transparent GPU memory deduplication framework for serverless LLM infe…
MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU Training
This paper presents MemX, a memory-aware automatic parallelism system for large-…
GPU Offload in Rust: Portable, Safe, and Fast
This paper presents a rustc-based GPU compilation framework for Rust, ensuring m…
TurboServe: Serving Streaming Video Generation Efficiently and Economically
TurboServe is a serving system designed for streaming video generation workloads…
ARGUS: Production-Scale Tracing and Performance Diagnosis for over 10,000-GPU Clusters
ARGUS is a low-overhead, fine-grained, always-on tracing and real-time analysis…
KineticSim: A Lightweight, High-Performance Execution Engine for Real-Time Market Simulators
Simulates financial markets at scale with parallel clearing pattern