Performance Optimization

This guide describes optimizations to achieve maximum streaming performance on Red Pitaya.


Overview

For maximum streaming performance, especially at high sample rates or with multiple channels, system-level optimizations are essential. The most impactful optimization is disabling the web interface, which frees up significant CPU resources and network bandwidth.

Note

These optimizations are particularly important when:

  • Streaming at sample rates above 10 MS/s

  • Using multiple channels simultaneously

  • Streaming over long durations

  • Running on lower-spec Red Pitaya models

  • Approaching the 62.5 MB/s network transfer limit


Quick Start: Disable Web Interface

The fastest way to optimize for streaming:

systemctl stop redpitaya_nginx
systemctl disable redpitaya_nginx

This immediately stops the web interface and prevents it from starting on next boot.

To restore the web interface later:

systemctl enable redpitaya_nginx
systemctl start redpitaya_nginx

Note

For complete service management instructions, including starting the SCPI server, automatic boot configuration, and troubleshooting, see the Service Management Guide.


Additional Optimizations

Memory Allocation

Red Pitaya boards have 512 MB of DDR3 RAM total. Ensure adequate Deep Memory Mode (DMM) allocation for streaming buffers while leaving sufficient memory for Linux operation:

  • Minimum: 100 MB for basic streaming

  • Recommended: 256 MB for high-rate streaming

  • Maximum: 400 MB (leaving at least 100 MB for Linux OS operation)

Note

Always reserve at least 100 MB of RAM for the Linux operating system. Allocating too much memory to streaming can cause system instability or crashes.

Configure DMM allocation in the web interface before disabling it, or manually via SSH:

nano /root/.config/redpitaya/apps/streaming/streaming_config.json

For more information on memory allocation and limits, see Deep Memory Mode.


Network Configuration

For maximum throughput:

  1. Use wired Ethernet (not Wi-Fi) - Gigabit Ethernet required for high-speed streaming

  2. Use router - Direct connection to PC is currently not supported for streaming

  3. Verify link speed is 1 Gbps:

    ethtool eth0

    Look for Speed: 1000Mb/s in the output. If the speed is lower (100 Mbps or 10 Mbps), check your Ethernet cable and switch/router capabilities.


System Load Reduction

Minimize background processes:

  1. Close unnecessary SSH connections (each consumes resources)

  2. Stop other Red Pitaya applications

  3. Disable the web interface when not needed (see Quick Start section above)

For more details, see Service Management.


Sample Rate Limits

Be aware of the maximum achievable rates:

  • Theoretical maximum: 125 MB/s (1 Gbps Ethernet)

  • Practical network limit: ~62.5 MB/s (achievable streaming rate)

  • Single channel: Up to 31.25 MS/s (62.5 MB/s ÷ 2 bytes/sample)

  • Dual channel: Up to 15.625 MS/s per channel (31.25 MS/s combined)

Note

The absolute maximum streaming rate cannot be tested due to Red Pitaya’s decimation limits. The practical limit is approximately 62.5 MB/s for sustained network streaming. Higher speeds may be possible with modifications to the FPGA firmware and custom software, but this is beyond the scope of typical usage (users have reporeted up to 100 MB/s with custom setups).

Going beyond these limits requires:

  • Reducing active channels

  • Using local SD card storage instead of network streaming

  • Increasing decimation (reducing sample rate)


Workflow Recommendations

Development Phase

During development and testing:

  1. Keep web interface enabled for easy configuration

  2. Use web interface to:

    • Configure streaming parameters

    • Set DMM memory allocation

    • Test basic functionality

  3. Monitor performance and identify bottlenecks


Production Phase

For production streaming or critical measurements:

  1. Configure all settings via web interface

  2. Download configuration using rpsa_client

  3. Disable web interface for maximum performance

  4. Verify streaming stability at target sample rates

  5. Use command-line client for all streaming operations

See CLI Examples for command-line workflows.


Verification

After optimization, verify performance:

Check Service Status

Confirm web interface is stopped:

systemctl status redpitaya_nginx

Expected output:

● redpitaya_nginx.service - Red Pitaya Nginx Web Server
     Active: inactive (dead)

Test Streaming Performance

Run a test capture at your target sample rate:

rpsa_client -h <red_pitaya_ip> -p TCP -f ./ -t wav

Monitor system resources during streaming:

top

Look for:

  • CPU usage <80% during streaming

  • No memory swapping

  • Stable network throughput


Troubleshooting

Data Loss Despite Optimization

If experiencing data loss after optimization:

  1. Reduce sample rate - May be exceeding network bandwidth

  2. Check decimation settings - Ensure correct calculation:

    Sample rate = 125 MS/s ÷ decimation
Network usage = sample_rate × channels × 2 bytes

  3. Verify network quality - Run ethtool eth0 to check link

  4. Increase DMM memory - Provides more buffering

  5. Check for packet loss:

    netstat -s | grep -i "packet.*loss"

Can’t Re-enable Web Interface

If web interface won’t restart:

  1. Check service status:

    systemctl status redpitaya_nginx

  2. Look for errors in logs:

    journalctl -u redpitaya_nginx -n 50

  3. Restart the service:

    systemctl restart redpitaya_nginx

  4. Reboot if necessary:

    reboot

For more troubleshooting, see Service Management.


Performance Still Poor

If performance doesn’t improve:

  1. Check hardware capabilities - Older boards may have limitations

  2. Verify network infrastructure - Test with different switch/cable

  3. Update Red Pitaya OS - Newer versions have performance improvements

  4. Review configuration - Incorrect settings can limit performance

  5. Check for thermal throttling - Ensure adequate cooling

See Streaming Limitations for detailed performance boundaries.


Additional Resources