Dramatically Accelerate 96Board Software via an FPGA with Integrated Processors
summary
Claims that a Zynq UltraScale+ based Ultra96 (ZU3) board plus FPGA fabric and embedded ARM cores let developers convert hot functions with the SDSoC HLS flow into DMA‑attached accelerators, yielding typical speedups like 10× FFT, 25× optical‑flow/stereo and an 8,600×–19,000× neural‑network boost on larger silicon, while cutting watts to single‑digit levels; shows a working CNN for German traffic signs (tiling → ~94 fps on Ultra96), published reference designs and a one‑year SDSoC license.
AI-written summary · about
chapters (9)
- 0:00Introduction and overview
- 1:08What is an FPGA?
- 8:35Heterogeneous all-programmable devices
- 16:54Custom peripherals and acceleration
- 28:20Software development flow
- 32:25Ultra 96 development platform
- 37:21Software acceleration techniques
- 43:27Neural network acceleration
- 54:08Wrap-up and resources
related talks
51mDesign your own custom co-processors and acceleration hardw
17mAccelerating Designs for SoC FPGA Using Simplified High-Level Synthesis Flows
45mAutomated, Simulation-Based Flow for Low-Cost FPGA-Accelerated Devices with Z...
49mFPGA: creating hardware by programming
25mSecuring your next 96Boards design using Xilinx Zynq MPSoC
43mSoftware Development Trends for FPGAs
27mGNU Radio in a Direct-RF World
25mOpen Source Software Stack for Heterogenous embedded devices