Modern Industrial Manufacturing
Industrial manufacturing involves complex processes that require extreme precision, efficiency, and reliability to maintain high productivity levels. High-Performance Computing (HPC) systems have become integral in this sector, providing the computational power needed for tasks ranging from product design and simulation to automation and real-time quality control. These systems enable manufacturers to achieve faster processing times, higher quality outputs, and more efficient use of resources.
Challenges in Industrial Manufacturing
Operating in an industrial setting introduces specific challenges for HPC deployments. The appliances each system is deployed in must achieve maximum performance while adhering to strict environmental standards to ensure safety and operational continuity. For instance, in regions prone to seismic activities, HPC systems need sturdy structural designs to withstand potential earthquakes. Additionally, the noise generated by these systems must be carefully managed to ensure it does not exceed safe levels, protecting the health and safety of workers on the manufacturing floor.
The need for customized HPC solutions in industrial manufacturing is driven by these unique environmental and operational requirements, highlighting the importance of specialized design and engineering to meet the specific needs of each manufacturing facility.
Project Scope and Client Needs
The primary objective for this project was to engineer a custom HPC manufacturing appliance that met highly specific operational requirements at scale. Given the industrial environment, the system needed to:
- Ensure Optimal Thermal Management: Achieve ultra-low CPU temperatures with a multi-stage liquid cooling system capable of 100% heat removal, critical for maintaining system performance and longevity under continuous operation.
- Mitigate Noise Pollution: Operate within stringent noise regulations to protect the hearing and well-being of workers on the production floor. This involved designing the cabinet to function at full load below 65dB (the noise level of a normal conversation).
- Enhance Structural Integrity: Incorporate a seismic cabinet design to ensure the system's resilience and stable operation in geologically active zones, preventing disruptions and damage during seismic events.
- Maintain Safety Standards: Include features such as an active leak detection system with a built-in leak tray and an emergency OFF button, ensuring rapid response capabilities and enhanced safety for onsite personnel.
AMAX's Customized HPC Solution
64-Node HPC Compute with IB Network
This core configuration was chosen for its ability to deliver substantial computational power necessary for industrial manufacturing processes. The InfiniBand network ensures high-speed connectivity and data transfer rates, critical for minimizing processing delays and enhancing overall system efficiency.
Advanced Multi-Stage Liquid Cooling System
To maintain the CPUs at ultra-low temperatures and manage 100% heat removal efficiently, AMAX implemented a sophisticated liquid cooling system. This system was designed for optimal thermal management and quite operations, reducing the impact on the ambient noise level on the manufacturing floor.
- Cold Plate Liquid Cooling: This direct contact method effectively draws heat away from the chips, significantly reducing the risk of overheating even under substantial computational loads. The cooling mechanism extends to the node level, maintaining a uniformly low temperature across all computing units.
- Rear Door Heat Exchanger (RDHx): Complementing the cold plate technology is the Rear Door Heat Exchanger (RDHx), a system that attaches to the rear of the computing racks. The RDHx works by regulating airflow, capturing and eliminating any residual heat before it enters the facility's ambient environment.
Seismic Cabinet Design
Recognizing the need for stability in a seismically active region, the HPC system was encased in a specially engineered cabinet. This design focused on mitigating vibrations and enhancing the structural integrity of the system, ensuring continuous operation during and after seismic events.
Integrated Safety Features
- Active Leak Detection System: To prevent and swiftly manage potential liquid cooling leaks, an advanced detection system was installed, complete with a built-in leak tray for immediate containment.
- Emergency Off Button: Strategically placed for accessibility, this feature provided a quick-response mechanism to immediately cease all operations, safeguarding both the system and the personnel in case of an emergency.
- Touchscreen Monitor: A front-facing touchscreen control panel enables monitoring of power distribution at PDU outlets, temperature, cooling systems, leakage detection, and computing performance.
Noise Reduction Techniques
Special attention was given to the design of components and the assembly of the system to ensure it operated under 65dB at full load. This not only helped in complying with workplace safety regulations but also contributed to a more conducive working environment for the floor staff.
Compliance Standards
Ensuring compliance with industry standards is fundamental in the development and deployment of industrial manufacturing equipment. For the customized HPC solution provided by AMAX, adherence to multiple global and regional standards was a critical aspect of the project, ensuring the system met the rigorous safety, quality, and environmental benchmarks required for industrial applications.
- SEMI S2/S8: These guidelines are crucial for safety in semiconductor manufacturing, encompassing a broad spectrum of equipment safety standards, including environmental, health, and safety (EHS) aspects, emergency shutdown systems, and ergonomic considerations for operator safety.
- NFPA 79: This standard provides guidelines for electrical safety in industrial machinery, ensuring that all electrical components and systems are designed to prevent failures and hazards, thus safeguarding both the equipment and its operators.
- Sil-3: The Sil-e standard refers to safety integrity levels that are crucial in ensuring that systems perform to the required safety levels throughout their service life.
- Cabinet-Level Safety: This involves comprehensive monitoring and control systems within the cabinet to ensure continuous safety and operational integrity, including aspects like temperature control and leak detection.
- Emissions Compliance: Meeting international standards such as FCC (USA), UL (USA), CE (Europe), KCC (Korea), and BSMI (Taiwan) ensures that the equipment adheres to required norms for electromagnetic compatibility and safety.
Serviceability and Reliability
To minimize downtime and facilitate maintenance, the system featured front hot-swappable SSDs and complete nodes that could be replaced quickly and easily. This approach not only enhanced the longevity of the system but also ensured consistent performance and reliability. Given the industrial manufacturing industry's practice of utilizing a single system design for several years, the durability and ease of maintenance of the system are critical. AMAX's design ensures that the system can continue performing optimally over its extended service life, aligning with the industry’s need for long-lasting and reliable solutions.
Advancing Production Through Precision Engineering
The deployment of AMAX's customized industrial HPC manufacturing appliance significantly enhanced the client’s production capabilities. By integrating a 64-node HPC compute cluster with an advanced IB network, the system drastically reduced processing times, enabling a marked increase in output efficiency. The multi-stage liquid cooling and stringent noise management further ensured that the system operated under optimal conditions, minimizing maintenance downtime and extending the system’s operational lifespan. As a result, the client reported a substantial improvement in overall productivity, affirming the value of AMAX’s tailored solutions in driving industrial manufacturing success.
Customized Solutions for Your Needs
Start designing your full-scale, custom HPC solution with AMAX today. We prioritize your unique requirements, engineering each solution from scratch to meet your specific needs. Our process is transparent and collaborative, ensuring that the final product enhances your operations performance and meets even the most stringent requirements.
Connect with AMAX to build a system that not only meets but exceeds your expectations.