A course on Accelerating Big Data Analytics Application with...
summary
Course explains how OpenCAPI FPGA engines accelerate big data by connecting FPGAs to host memory, uses Fletcher with Apache Arrow to build platform‑agnostic lab pipelines that run on laptops and scale to POWER systems, and teaches two concrete projects: high‑throughput x‑ray acquisition and genomic acceleration with ready GitHub labs and VM/docker images. Covers Power memory inception as a routing idea with ~2 TB/s aggregate bandwidth and cites Arrow activity ~50 commits/week.
AI-written summary · about
chapters (20)
- 0:00Introduction and presenters
- 1:04Course overview and agenda
- 2:11Why hardware acceleration for big data
- 4:40Evolution of accelerator technologies
- 7:50GPU vs FPGA comparison
- 9:33What is OpenCAPI
- 11:19Collaborative course development
- 12:45Course highlights and features
- 15:15How to access the course
- 18:04Motivation for big data analytics
- 19:08In-memory computing benefits
- 21:59Challenges of in-memory computing
- 23:49Addressing big data challenges
- 25:55Apache Arrow in-memory format
- 27:40Hardware standards for memory sharing
- 28:42Memory Inception and routing
- 33:38CPU bottlenecks and accelerators
- 36:25Course collaboration and Q&A
- 37:26Audience questions and answers
- 40:57Closing remarks and thanks
related talks
37mAccelerated Big Data Systems - Experiences with OpenCAPI
07mData analytics accelerated with FPGA's
33mArrow in Flight - New Developments in Data Connectivity
22mAccelerating Analytics with Apache Arrow
45mAccelerating OpenSearch With Streaming: Apache Arrow, Flight, Data...
33mA 101 in time series analytics with Apache Arrow, Pandas and Parquet
45mApache Arrow: In Pursuit of Speed
34mApache Arrow as a full stack data engineering solution