MSHGN: Multi-scenario adaptive hierarchical spatial graph convolution network for GPU utilization prediction in heterogeneous GPU clusters

Article Computer Science, Theory & Methods

Horus: Interference-Aware and Prediction-Based Scheduling in Deep Learning Systems

Gingfung Yeung et al.

Summary: In this article, we propose Horus, an interference-aware and prediction-based resource manager for DL systems, which achieves high resource utilization and reduces makespan by predicting GPU utilization.

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS (2022)

添加到收藏夹

Article Computer Science, Theory & Methods

Workload forecasting and energy state estimation in cloud data centres: ML-centric approach

Tahseen Khan et al.

Summary: Predicting workload and energy consumption in cloud data centers is crucial for efficient resource management. Using machine learning models and clustering algorithms, workload and energy state of virtual machines can be predicted to aid in automated resource management decisions.

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE (2022)

添加到收藏夹

Review Computer Science, Hardware & Architecture

Status, challenges and trends of data-intensive supercomputing

Jia Wei et al.

Summary: This paper provides an overview of the key concepts and developments in supercomputing, focusing on data-intensive computing. It discusses the challenges and demands of data-intensive supercomputing and explores the future trends and potential challenges in this field.

CCF TRANSACTIONS ON HIGH PERFORMANCE COMPUTING (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

An efficient forecasting approach for resource utilization in cloud data center using CNN-LSTM model

Soukaina Ouhame et al.

Summary: The proposed hybrid model improves the accuracy rate by approximately 3.8% to 10.9% and reduces the error percentage rate by approximately 7% to 8.5% compared to other models. This technique enhances central processing unit, memory, disk, and network usage in the network, providing better prediction results than other models.

NEURAL COMPUTING & APPLICATIONS (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Integrated deep learning method for workload and resource prediction cloud systems

Jing Bi et al.

Summary: Cloud computing providers face challenges in forecasting large-scale workload and resource time series. By using logarithmic operation, powerful filters, and deep learning methods, more accurate predictions can be achieved.

NEUROCOMPUTING (2021)

添加到收藏夹

Article Computer Science, Information Systems

BHyPreC: A Novel Bi-LSTM Based Hybrid Recurrent Neural Network Model to Predict the CPU Workload of Cloud Virtual Machine

Md Ebtidaul Karim et al.

Summary: With the advancement of cloud computing technology, the demand for maximizing cloud resources is increasing. Effective virtual machine consolidation and migration can reduce energy consumption in cloud data centers, while accurate workload prediction is crucial for effective task scheduling. Our proposed hybrid prediction model, BHyPreC, utilizing Bidirectional LSTM on top of stacked LSTM and GRU, demonstrates better accuracy compared to other statistical models for predicting cloud VM's future CPU usage workload.

IEEE ACCESS (2021)

添加到收藏夹

Article Computer Science, Information Systems