x86 Linux

x86_64 Linux

The x86_64 platform is primarily used for development, testing, and evaluation. It runs the same codebase as the embedded targets, making it ideal for:

• Rapid prototyping and algorithm evaluation

• CI/CD test pipelines

• Batch processing audio files

• Performance benchmarking

Setup

# Clone the repository
git clone <repo-url>
cd c

# Build for x86_64
make

# Run with a test file
./cry_detect -f test.wav -e pure_c -w model_weights.bin

Backend Options

NPU engines ("svp_acl", "magik") are not available on x86 — they require the respective hardware and SDK.

Development Workflow

1. Develop and test on x86 — use WAV files with cry_detect_process_file()

2. Validate accuracy — compare detection outputs against ground truth

3. Cross-compile for target — build for ARM/MIPS with the appropriate toolchain

4. Deploy and verify on target — run the same test files on the embedded device

Audio Capture

On x86 Linux, audio capture defaults to "file" mode. For microphone input during development, you can:

• Use `arecord` to capture WAV files, then process them with `cry_detect_process_file()`

• Contact us for ALSA-based capture backend support