Hardware Accelerated Big Data Processing
summary
Conference demonstrates concrete use of FPGAs for streaming big data with measurable results: a PCIe FPGA board implementation of FPGA-accelerated XML schema validation processes up to ~3+ Gbps versus ~300 Mbps on an Intel Xeon comparison, yielding roughly a 10x speedup. Description cites Microsoft’s Bing experiments and Intel’s Altera acquisition and proposes mapping automata to LUTs and flip-flops, uses on-board RAM and DSP blocks, and reports a working prototype with realistic server deployment and modest TCO/power gains.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:35FPGAs
- 2:25Microprocessors
- 5:55FPGA
- 8:00Logic Block
- 8:20Layout
- 9:40Programming
- 10:40Big Data
- 12:45Microsoft
- 13:50Theory
related talks
06mAchieve predictable performance
28mAchieving 10Gbps Line-rate Key-value Stores with FPGAs
23mEnabling SQL to FPGA Using HLS for Filtering/Stream Data Transformations
27mFPGA-Accelerated Compactions for LSM based Key-Value Store
48mAnalytics Architectures on Microsoft Fabric
28mCAPI/OpenCAPI Accelerated PostgreSQL
18mThe "New Normal" or An Evolution
26mReimagine Real-time