Beyond Hotness: High-Resolution Memory Characterization of Cloud and LLM Workloads for Tiered Memory with ARM SPE

Jie Ren; Ivy Peng

Home/Archive/arXiv

arXiv · 2610.32711

Beyond Hotness: High-Resolution Memory Characterization of Cloud and LLM Workloads for Tiered Memory with ARM SPE

Jie Ren·Ivy Peng

View original source ↗Read source document ↗Export metadata ↓

Abstract

Memory demands from data analytics and LLM workloads grow rapidly, driving up memory costs in cloud and data centers. Multi-tier memory has emerged as a promising solution to balance performance and cost. However, effectively leveraging it requires a deep understanding of workload memory behaviors. This work introduces a high-resolution memory characterization methodology for analyzing workload-specific opportunities for multi-tier memory usage. In response to the increasing deployment of ARM-based processors in data centers, we implement our methodology using the recent Statistical Profiling Extension (SPE) for ARM. By characterizing eight memory-intensive workloads from CloudSuite and LLM applications, we identify region-specific access patterns, evaluate the effectiveness of cloud-oriented cache hierarchies, and propose new metrics based on memory idle periods and latency profiles to determine suitable data placement for tiered memory.

Explore related subjects

cs.DC

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Jie Ren, Ivy Peng. 2026-10-23. Beyond Hotness: High-Resolution Memory Characterization of Cloud and LLM Workloads for Tiered Memory with ARM SPE. https://arxiv.org/abs/2610.32711

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

FastGraph: PCA-Subspace Binned k-Nearest-Neighbor Graph Construction for GPU Geometric Deep Learning

Dynamic graph neural networks repeatedly construct k-nearest-neighbor graphs in learned latent spaces, making GPU-accelerated graph construction an important operation in scientific point-cloud workloads. We present FastGraph, a PyTorch-integrated primitive for exact kNN graph construction in low-to-moderate-dimensional spaces. FastGraph uses a PCA-subspace cell list: it fits a compact binning coordinate system to the input tensor, uses that subspace for spatial pruning, and evaluates all candidate distances in the original feature space. The data-adaptive orthonormal subspace improves uniform-grid pruning in the d=4-10 range evaluated around the d=4 GravNet operating point. Point data remain GPU-resident, with small host synchronizations for control metadata. The implementation ships with an eager-mode PCA API and an opt-in differentiable GravNetOp integration. On a 5 M-point, single-segment detector-derived HGCAL recHit stress test, FastGraph builds the exact graph in 7.5 s at d=8, k=40, achieving speedups of up to 41x over FAISS-GPU, 19x over cuVS brute force, and 17x over approximate CAGRA while retaining exact neighbor sets under the paper's distance-based convention.

cs.DC↗

arXiv2026

Resource-Aware Task Allocator Design: Insights and Recommendations for Distributed Satellite Constellations

We present the design of a Resource-Aware Task Allocator (RATA) and an empirical analysis in handling real-time tasks for processing on Distributed Satellite Systems (DSS). We consider task processing performance across low Earth orbit (LEO) to Low-Medium Earth Orbit (Low-MEO) constellation sizes, under varying traffic loads. Using Single-Level Tree Network(SLTN)-based cooperative task allocation architecture, we attempt to evaluate some key performance metrics - blocking probabilities, response times, energy consumption, and resource utilization across several tens of thousands of tasks per experiment. Our resource-conscious RATA monitors key parameters such as arrival rate, resources (on-board compute, storage, bandwidth, battery) availability, satellite eclipses' influence in processing and communications. This study is an important step towards analyzing the performance under lighter to stress inducing levels of compute intense workloads to test the ultimate performance limits under the combined influence of the above-mentioned factors. Results show pronounced non-linear scaling: while capacity increases with constellation size, blocking and delay grow rapidly, whereas energy remains resilient under solar-aware scheduling. The analysis identifies a practical satellite-count limit for baseline SLTNs and demonstrates that CPU availability, rather than energy, is the primary cause of blocking. These findings provide quantitative guidance by identifying thresholds at which system performance shifts from graceful degradation to collapse.

cs.DC↗

arXiv2026

From a Last-Week Add-On to First-Day Practice: Three Years of Integrating MPI Performance Analysis into HPC Education with EduMPI Suite

Performance analysis is essential for scalable MPI programs, yet tool and workflow complexity often relegates it to the end of Parallel Programming courses. EduMPI Suite lowers this entry barrier by automating cluster execution, measurement, and near-real-time analysis while visualizing MPI communication from the first program. This paper presents a three-year experience report on its integration into HPC education. We trace the transition of performance analysis from a last-week add-on to a recurring element of MPI development, using unpublished historical survey evidence and individual 2025 questionnaire data. Earlier cohorts perceived performance analysis as unnecessary, difficult, or unintuitive. In the final study, all students recommended introducing EduMPI Suite early in the course. Of 33 students, 29-33 rated it higher than conventional tools for understanding program behavior, interpreting communication, identifying bottlenecks, obtaining visual feedback, and navigating the tool. These findings support embedding performance analysis throughout MPI education before transitioning learners to professional environments.

cs.DC↗