Skip to main content

Hardware Acceleration

The proxy includes automatic GPU detection and hardware-accelerated transcoding via FFmpeg. When a compatible GPU is available, encoding is offloaded from the CPU — dramatically increasing throughput and reducing latency.

warning

Hardware acceleration is only available when using as an external instance. Using the AIO container with the embedded proxy instance does not support hardware acceleration.

info

Hardware acceleration only applies to transcoding. Pass-through streaming (the default) does not use the GPU.

Supported Hardware​

NVIDIA GPUs​

  • Requirements: NVIDIA Container Toolkit
  • Acceleration: CUDA, NVENC, NVDEC
  • Performance: 10–20× faster than CPU encoding
  • Best for: High-concurrency setups with 4+ simultaneous transcoded streams

Intel GPUs​

  • Requirements: /dev/dri devices passed to the container
  • Acceleration: VAAPI, QuickSync (QSV)
  • Performance: 3–8× faster than CPU encoding
  • Best for: Efficient transcoding with good quality/performance balance

AMD GPUs​

  • Requirements: /dev/dri devices passed to the container
  • Acceleration: VAAPI
  • Performance: 3–5× faster than CPU encoding
  • Best for: Open-source acceleration on AMD hardware

CPU Fallback​

When no GPU is detected, the proxy automatically falls back to software encoding (libx264, libx265). Single stream recommended for CPU-only setups.

Docker Setup​

NVIDIA GPU​

Install NVIDIA Container Toolkit on the host first.

services:
  m3u-proxy:
    image: sparkison/m3u-proxy:latest
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    ports:
      - "8085:8085"

Or with docker run:

docker run -d --name m3u-proxy \
  --gpus all \
  -p 8085:8085 \
  sparkison/m3u-proxy:latest

Intel / AMD GPU​

services:
  m3u-proxy:
    image: sparkison/m3u-proxy:latest
    devices:
      - /dev/dri:/dev/dri
    ports:
      - "8085:8085"

Or with docker run:

docker run -d --name m3u-proxy \
  --device /dev/dri:/dev/dri \
  -p 8085:8085 \
  sparkison/m3u-proxy:latest

CPU Only​

No special configuration needed:

docker run -d --name m3u-proxy \
  -p 8085:8085 \
  sparkison/m3u-proxy:latest

Auto-Detection​

On startup, the container automatically:

  1. Detects available GPU hardware using lspci
  2. Checks device accessibility (/dev/nvidia*, /dev/dri/*)
  3. Tests FFmpeg capabilities for the detected hardware
  4. Selects the optimal encoder (NVENC, VAAPI, or software)
  5. Logs the result

Example startup output:

📦 m3u-proxy starting up...
🎬 FFmpeg version: 8.0
🔍 Running hardware acceleration check...
🔍 Hardware detection: NVIDIA GPU (GeForce RTX 3080)
✅ FFmpeg NVIDIA acceleration: AVAILABLE
🚀 Starting m3u-proxy application...

The detected acceleration type is available via the environment variables HW_ACCEL_AVAILABLE, HW_ACCEL_TYPE, and HW_ACCEL_DEVICE.

Performance Comparison​

HardwareConcurrent streamsCPU usageEncoding speed
NVIDIA GPU4+Very low10–20× CPU
Intel/AMD GPU1–2Low3–8× CPU
CPU only1HighBaseline

Troubleshooting​

No GPU detected

# Check GPU device availability in container
docker run --rm sparkison/m3u-proxy:latest ls -la /dev/dri /dev/nvidia* 2>/dev/null

# Check detection logs
docker logs m3u-proxy | grep -E "🔍|✅|❌"

NVIDIA issues

# Test NVIDIA container runtime on the host
docker run --rm --gpus all nvidia/cuda:12.0-base-ubuntu22.04 nvidia-smi

# Verify host driver
nvidia-smi

Intel/AMD issues

# Check DRI devices on the host
ls -la /dev/dri/

# Test VAAPI inside the container
docker run --rm --device /dev/dri:/dev/dri sparkison/m3u-proxy:latest vainfo

Verifying GPU Access​

After starting the container:

docker exec -it m3u-proxy ls -la /dev/dri      # Intel/AMD
docker exec -it m3u-proxy nvidia-smi           # NVIDIA

Next Steps​

Once hardware acceleration is configured, see Transcoding for how to create transcoded streams and customise encoding profiles.