Radio-ML ConvNet: Throughput - Radio-ML ConvNet: Throughput - 2025.2 English - XD100

Vitis Tutorials: AI Engine Development (XD100)

Document ID
XD100
Release Date
2026-03-27
Version
2025.2 English

From the following figure, calculate throughput based on AI Engine simulations. Eight inferences run in 532.8 μs, which translates to ~15,000 inferences per second.

figure