Development Environment Setup

Cross-compilation Docker Environment (x86 Host)

Used to compile arm64 applications on an x86 host for execution on A8625MY1.

Environment Information

Item	Path
SYSROOT root directory	/usr/local/rb-debug-x86_64/sysroots/aarch64-oe-linux/
Header directory	/usr/local/rb-debug-x86_64/sysroots/aarch64-oe-linux/usr/local/include
Library directory	/usr/local/rb-debug-x86_64/sysroots/aarch64-oe-linux/usr/local/lib

Standard Build Workflow

# Enter sample project directory
cd /opt/aplux/aidlite_qnn236_le_example
# Create build directory and compile
mkdir -p build && cd build
cmake -DCMAKE_TOOLCHAIN_FILE=/opt/aplux/aidlite_qnn236_le_example/toolchain.cmake ..
make -j $(nproc)
# Deploy to A8625MY1 target device
scp qnn_yolov5_multi aidlux@192.168.100.3:/usr/local/share/aidlite/examples/aidlite_qnn236/cpp/

Local Runtime Environment (A8625MY1 Compute Card)

System Access Methods

A8625MY1 is preinstalled with native Ubuntu 22.04 and supports two access methods:

1. ADB command-line access

# Check device connection on Windows/Linux
adb devices
# Enter device shell (root by default)
adb shell

2. SSH remote access

# Ensure the compute card network is configured and SSH is enabled
ssh aidlux@<Device_IP>
# Default username/password: aidlux/aidlux

Runtime Notes

Due to system libstdc++ version constraints, all AI programs must preload the specified library version at runtime:

LD_PRELOAD=/usr/lib/libstdc++.so.6.0.30 ./your_program