PERFORMANCE_SNAPSHOT.md

CHaiDNN-v2
Analysis and Eval
	Supported Layers		Performance/Resource Utilization
	Performance Eval
Design and Development
	API Reference	Quantization User Guide for CHaiDNN	Model Zoo	Running Inference on new Network
	Creating SDx GUI Project	Configurable Parameters	Custom Platform Generation	Software Layer Plugin
	SDSoC Environment User Guide		Hardware-Software Partitioning for Performance

Performance Snapshot

The following table lists the accuracy of networks quantized for inference at different bit-widths.

• Caffe with Xilinx Quantization refers to simulation of Xilinx quantizer using Caffe framework.
• CHaiDNN with Xilinx Quantization refers to inference of Xilinx quantized network using CHaiDNN accelerator.

The following table lists the performance numbers on ZU9.

The following table lists the performance numbers on ZU7 (As resources in ZU7 are lower than ZU9, standalone pool kernel is disabled and the pool functionality in Convolution kernel is used).

The following table lists the performance numbers for DietChai and DietChai Plus.

^* Only accelerator resource utilization numbers are included.

---

^{Copyright© 2018 Xilinx}