2023 is over, so we’ve compiled a collection of 2023’s most popular content according to our readers. In case you missed anything, here’s your chance to catch up on best practices ebooks, technical blogs, hands-on videos, webinars and more.
Enjoy!
ALL THINGS AI
Building High-performance Data Access Layer for Model Training and Model Serving for LLM , User Blog
Mengyu Hu and Chengkun Jia, both from Zhihu’s data platform team, discuss their evolution from HDFS to Alluxio as a high-performance data access layer for LLM training and serving. Alluxio has accelerated model training by 2~3x, increased GPU utilization to 90%, and enabled model deployment every minute instead of hours or days.
Efficient Data Access Strategies For Large-scale AI , Whitepaper
Get a comprehensive understanding of data access patterns in a modern AI/ML platform. This white paper discusses the characteristics of data access in each stage of the machine learning pipeline and the solutions that can be used in architecting your data and AI platform.
Maximize GPU Utilization for Model Training , On-demand Webinar
When training models on ultra-large datasets, one of the biggest challenges is low GPU utilization. These powerful processors are often underutilized due to inefficient I/O and data access. This mismatch between computation and storage leads to wasted GPU resources, low performance, and high cloud storage costs. The rise of generative AI and GPU scarcity is only making this problem worse.
In this webinar, Tarik and Beinan discuss strategies for transforming idle GPUs into optimal powerhouses. They will focus on cost-effective management of ultra-large datasets for AI and analytics.
End-to-End Machine Learning Pipeline with Alluxio , 3min Demo
Watch the Alluxio Enterprise AI end-to-end ML pipeline demo, and see for yourself the significant performance improvements as well as increased GPU utilization! Alluxio’s Solution Engineer Tarik Bennett walks through a short end-to-end machine learning pipeline with Alluxio provisioned or mounted as a local folder for PyTorch dataloader.
Solving the Data Loading Challenge for Machine Learning with Alluxio , 3min Video Series
Alluxio’s Senior Solutions Engineer Roland Theron shares how Alluxio benefits model training workflows by reducing data loading times, allowing for better utilization of your compute resources.
Discover the easily consumed tuning tips that deliver optimal training speeds at lower costs. Learn how to tune PyTorch performance to achieve lower latency and higher GPU utilization through data loading, data operations, GPU processing, and CPU processing, with lines of code.
Rise of the Data Access Layer for Analytics & AI , Analyst Research
Explore the transformative capabilities of the Data Access Layer and how it can simplify and accelerate your analytics and AI workloads. in this new research paper, Kevin Petrie, VP of Research at Eckerson Group, shares the architecture and use cases for a Data Access Layer and how it can help achieve analytics and AI goals with successful performance.
COST SAVINGS AND OPTIMIZATIONS
Millions Saved Annually: Unleashing the Power of Alluxio + HDFS at Uber , User Blog
Find out details of our joint project with Uber aimed at optimizing the performance of HDFS DataNodes. The project utilized the Alluxio SDK cache to manage an SSD storage on each DataNode, resulting in improved performance and a better return on investment. Despite the SSD cache occupying only 0.6% of the total disk space, it impressively handles 60% of the overall client traffic.
The Ultimate Guide to Saving Data Egress Costs in the Cloud , Ebook
Build your data platform with reduced cloud egress costs and never be surprised by a bill again. Minimize your data replication with optimized data pipelines and data flow for your architecture.
Shopee: Query Acceleration & Data Access as a Service , User Blog
Learn how Shopee, the leading e-commerce platform in Asia, has successfully leveraged Alluxio to improve Trino query performance by ~55%. In addition, Alluxio enhances developer experience by providing flexible data access through Data APIs.
The Trino Optimization Handbook , Ebook , Best Practices and Tuning Tips
Unlock the full potential of Trino and transform your data analytics game. Identify bottlenecks and maximize your Trino query performance with configuration settings and session properties.
If you’re using Presto (PrestoDB), check out The Presto Optimization Handbook here.
(image source: bhargavkesavan.wordpress.com)
Happy reading! If you have any questions, chat with us on Slack!
.png)
Blog
We are thrilled to announce the general availability of Alluxio Enterprise for Data Analytics 3.2! With data volumes continuing to grow at exponential rates, data platform teams face challenges in maintaining query performance, managing infrastructure costs, and ensuring scalability. This latest version of Alluxio addresses these challenges head-on with groundbreaking improvements in scalability, performance, and cost-efficiency.
.jpeg)
We’re excited to introduce Rapid Alluxio Deployer (RAD) on AWS, which allows you to experience the performance benefits of Alluxio in less than 30 minutes. RAD is designed with a split-plane architecture, which ensures that your data remains secure within your AWS environment, giving you peace of mind while leveraging Alluxio’s capabilities.
PyTorch is one of the most popular deep learning frameworks in production today. As models become increasingly complex and dataset sizes grow, optimizing model training performance becomes crucial to reduce training times and improve productivity.