> For the complete documentation index, see [llms.txt](https://acpcoin.gitbook.io/acp-coin-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://acpcoin.gitbook.io/acp-coin-docs/2.1-prime-mining-architecture.md).

# 2.1 Prime Mining Architecture

Prime Mining operates under the RSR system, enabling participants to connect their GPUs to the network, process AI tasks, and receive real-time rewards. It does more than share resources. It creates value from industrial-level computing that is fed back into the network.

**2.1.1 System Overview**

* **Participant Nodes (Clients)**: Individuals or organizations that join the network by connecting GPU hardware via the Prime Mining client.<br>
* **Orchestrator Nodes**: Entities that manage task distribution and verify results, interfacing directly with smart contracts to finalize payments.<br>
* **Computation Consumers**: AI API users, enterprises employing LLMs, and other service platforms that request compute power directly through the RSR network.\ <br>

**2.1.2 Operation Flow**

1. **Node Registration** – Users install the Prime Mining client and register their GPU device.<br>
2. **Task Allocation** – Orchestrators distribute incoming AI compute requests across participant nodes.<br>
3. **Execution & Submission** – Each node processes the assigned task and submits output hashes to the smart contract.<br>
4. **Verification & Reward** – Once computation is verified, rewards are issued to contributors based on performance.\ <br>

**2.1.3 Technical Highlights**

* **Distributed Parallelization**: Tasks are broken down into smaller units and processed simultaneously across nodes, significantly improving speed and efficiency.<br>
* **Multi-GPU Optimization**: Algorithms dynamically allocate workloads across GPUs to maximize performance and reward precision.<br>
* **Real-time Node Monitoring**: Node availability, accuracy, and stability are continuously tracked. Oracle services and digital signatures safeguard data integrity.<br>

AI-Task-Aware Routing: Jobs are routed according to purpose (e.g., LLM inference, image generation, data prep), ensuring optimal node-task matching.