English

Performance Indicator

This section provides you with instructions on how to use the HPL performance data testing platform and shows you the performance data result statistics for reference.

Test Platform

Test conditions (hardware and software platform requirements):

Test development board: J6 development board.
OpenCV version: 3.4.5.
OpenCV runtime environment: A78 single-threaded.

Test Method

Performance test method: Since the time consumed to run the operator will fluctuate, here we take the average performance data of multi runs of the operator.
Test time with steady_clock.

std::chrono::duration_cast<std::chrono::microseconds>(std::chrono::steady_clock::now().time_since_epoch()).count();

Data description:
- Task time: Operator task execution time.
- Baseline Time: OpenCV execution time.
- Ratio: Baseline Avg / Task Avg.

Performance Comparison

Note

The number of FFT points used in the test is 1024 points, and the amount of data used in the FFT 1D test is 20,480,000 floating point numbers.

Algorithm	High performance library(us)	OpenCV 3.4.5 A78(us)	Ratio(default DSP)
fft1d	98576	1109059	11.251
ifft1d	98628	1149626	11.656

On This Page

The hb_compile Tool

QAT

Model Export

Horizon Operator

Vision Algorithm

Sample

VP API Introduction

Data Structure

Function Interface

Model Inference API Introduction

Data Type and Data Structure

API Interface

Model Inference Tool Introduction

High Performance Algorithm

HPL API Introduction

Data Structure

Functional Interface

Data Structure

Functional Interface

FRAMEWORK

TUTORIALS

EXAMPLES

API REFERENCE

#Performance Indicator

#Test Platform

#Test Method

#Performance Comparison

Performance Indicator

Test Platform

Test Method

Performance Comparison