Provider Guide

Join DCP as a GPU Provider

List your NVIDIA GPU on the Saudi DCP marketplace and begin receiving suitable jobs.

Requirements

NVIDIA GPU

NVIDIA GPU with at least 8 GB VRAM (16 GB+ recommended for stronger matching).

Operating System

Windows 10/11 or Ubuntu 20.04/22.04/24.04. NVIDIA drivers must be installed.

Internet Connection

100 Mbps symmetric or faster internet is recommended for stable workload execution.

DCP Account

Python 3.8+, 50 GB free disk, Docker, and NVIDIA Container Toolkit are required in the provider runtime stack.

Earnings

You keep 75% of every settled compute job. Earnings are posted to your dashboard after job completion.

75% Revenue Split

DCP takes a 25% platform fee. The remaining 75% is credited to your wallet when completed jobs settle.

RTX 3090 Example

Use your utilization and demand assumptions to build planning scenarios. Results vary by job mix, uptime, and active matching.

Withdrawal Requests

Submit a withdrawal request from your dashboard to move earnings from your wallet through the payout workflow.

Setup in 6 Steps

1. Check Prerequisites

Verify your system: NVIDIA GPU with CUDA support, Ubuntu 20.04+ or Windows 10/11, Python 3.8+, and Docker (recommended for job isolation).

# Verify GPU driver
nvidia-smi

# Verify Python 3.8+
python3 --version

# Verify Docker (recommended for job isolation)
docker --version

2. Create Your Provider Account

3. Download the Daemon

Download dc1_daemon.py with your API key pre-injected. Replace YOUR_PROVIDER_KEY with the key from step 2. The file is personalized — do not share it.

# Linux / macOS — downloads daemon with your key pre-injected
curl "https://dcp.sa/api/dc1/providers/download/daemon?key=YOUR_PROVIDER_KEY" \
  -o dc1_daemon.py

# Make executable (Linux)
chmod +x dc1_daemon.py

4. Run the Daemon

Start the daemon process. It will detect your GPU, register with DCP, and send status heartbeats. Keep the process running or run it as a service.

# Linux / macOS
python3 dc1_daemon.py

Expected output

[DCP] dc1_daemon v3.3.0 starting
[DCP] GPU detected: NVIDIA RTX 4090 (24 GB VRAM)
[DCP] Provider ID: 42 | Status: online
[DCP] Heartbeat sent — next in 30s
[DCP] Waiting for jobs...

5. Verify in Dashboard

Open dcp.sa/provider and confirm your GPU is listed as online. You can also verify via the API with your provider key.

# Check your provider status via API
curl "https://dcp.sa/api/dc1/providers/me?key=YOUR_PROVIDER_KEY"

Expected output

{
  "provider": {
    "id": 42,
    "status": "online",
    "gpu_model": "RTX 4090",
    "last_heartbeat": "2026-03-19T18:00:00Z",
    "total_jobs": 0,
    "total_earned_sar": "0.00"
  }
}

6. Verify matched workload processing

When routing places a compatible workload, the daemon runs it in an isolated Docker container. Completed jobs create wallet entries in your provider dashboard.

Expected output

[DCP] Job received: job-abc123 (llm_inference, 30 min)
[DCP] Pulling Docker image: vllm/vllm-openai:latest
[DCP] Container started — executing job
[DCP] Job completed in 612s
[DCP] Earnings credited: +45.75 SAR (75% of 61.00 SAR)

Onboarding status troubleshooting matrix

Each status has one deterministic next action. Use the guide for commands or open support with this state prefilled.

Registered (waiting for first heartbeat)

Run the install command once, keep the daemon process open, and wait for the first heartbeat.

python3 dc1_daemon.py

Heartbeat received (final checks in progress)

Keep the daemon running and verify recent heartbeat timestamps before waiting for jobs.

curl "https://dcp.sa/api/dc1/providers/me?key=YOUR_PROVIDER_KEY"

Heartbeat stale or disconnected

Restart the daemon, then verify API key validity and endpoint reachability checks.

curl -I "https://dcp.sa/api/dc1/providers/download/daemon?key=YOUR_PROVIDER_KEY"

Paused

Resume your provider from the dashboard and confirm heartbeat polling resumes.

Open /provider/dashboard and select Resume

Ready for matched jobs

Keep the daemon online and monitor your dashboard for queued compute jobs.

Open /provider/dashboard and check jobs + earnings panels

Frequently Asked Questions

Ready to Earn with Your GPU?

Publish compatible capacity in the Saudi GPU marketplace and start workload routing.