Understanding the Open Architecture of AI Hardware: OAI ＆ OAM
In the AI domain, the Open Accelerator Infrastructure (OAI) is a sub-organization established by one of the world's most influential open hardware organizations: the Open Compute Project (OCP). Since 2019, OAI has focused on defining AI accelerator card form factors suitable for large-scale deep learning training, addressing the issue of diverse AI accelerator card forms and interface incompatibility. By releasing the OAI-UBB (Universal Base Board) 1.0 design specification, OAI has promoted the standardization of AI acceleration hardware platforms, enabling support for products from different manufacturers without modification, significantly enhancing the scalability and flexibility of AI modules.
The Architecture Adopted by AI Giants - OAM (Open Accelerator Module)
For AI server developers, the Open Accelerator Module (OAM) brings significant benefits. Due to the diversity and specialization of AI accelerator chips, developers face higher development costs and longer development cycles. The emergence of OAM provides these developers with an efficient and scalable solution, making it easier to integrate new AI accelerators. This not only lowers the entry barrier but also accelerates the time-to-market for products.
Advantages and Challenges of OAM
OAM has three notable advantages, making it particularly important in today's rapidly developing AI market:
High Performance and Efficiency: OAM can significantly improve processing performance and efficiency, particularly for high-computing-demand applications such as deep learning and machine learning.
Scalability: OAM's design allows for flexible compatibility and scalability between different systems and baseboards, adapting to the advancing high-computing-load and technological developments.
Support for Diverse Application Scenarios: OAM is applicable to various fields, including AI inference, scientific simulations, and data analysis, enabling it to meet a wide range of business needs.
These advantages demonstrate OAM's potential and flexibility in modern data centers and high-performance computing environments.
However, adopting OAM also comes with several challenges:
Technical and Design Complexity: Current specialized AI hardware systems are technically and design-wise complex, often requiring 6 to 12 months to integrate new AI accelerators into the system, thus hindering the rapid adoption of new competitive technologies.
High Power Consumption Management: With OAM product design power exceeding 600W, the Base Specification already recommends the use of liquid cooling technology, indicating that managing high power consumption is a significant challenge, especially as future power may easily exceed 700W.
Standardization and Compatibility: The rapid evolution and diversity of AI accelerators demand higher standards for standardization and compatibility to support the scalability and high-speed communication links between various hardware acceleration solutions.
What Can NEXTRON Do for OAM Designers?
NEXTRON is one of the few market solution providers with expertise in high-speed transmission, structural design, and cooling modules. It has already provided high-speed I/O and OAM-related products to several leading AI chip designers. For OAM design, NEXTRON has identified two common challenges:
Structural Design for Cooling: Top Stiffener with Thermal Solution The Top Stiffener, though minimally covered in the Base Specification, requires good structural design to complement the OAM design; otherwise, it directly impacts cooling efficiency. A good Top Stiffener design includes solid support and thermal conduction, enabling air-cooling solutions like 3U or 4U height 3D VC to function effectively. Facing next-generation solutions with over 600W TDP, integration with a cold plate liquid cooling is also necessary. This tests the manufacturer's structural design, material selection, and understanding of cooling technology.
Challenges in Processing: OAM Bottom Stiffener Poor connection between OAM and UBB is a common issue in practice, often due to neglecting the OAM Bottom Stiffener. OAM generally uses the Mirror Mezz Pro Connector, and the Base Specification mentions that tolerances must be maintained at ±0.15mm. However, due to differences in assembly and processing capabilities, a poorly made Bottom Stiffener can lead to excessive final assembly tolerances or uneven ends, resulting in poor connections.
Conclusion
AI-related technologies have developed rapidly in recent years, bringing cross-disciplinary challenges. As NEXTRON collaborates with AI leaders to overcome various difficulties, it deeply realizes that every small detail is crucial for achieving impressive AI performance. NEXTRON hopes that the accumulated experiences and capabilities can help more AI product developers solve problems, making this civilization-changing technology more accessible to everyone.
- +1 Like
- Add to Favorites
This document is provided by Sekorm Platform for VIP exclusive service. The copyright is owned by Sekorm. Without authorization, any medias, websites or individual are not allowed to reprint. When authorizing the reprint, the link of www.sekorm.com must be indicated.
Recommend
Nextron’s Next Generation QSFP112 cage/connector system and Heat Dissipation Strategies
2022-09-02 - Technical Discussion In this article we discuss 3 most popular types of thermal management, air cooling methods, liquid cooling systems, thermal conductive material and additions. Then Nextron tells the heat dissipation solutions based on QSFP112.
Building AI Servers with DRAM for Artificial Intelligence
2023-08-18 - Technical Discussion ATP Electronics meets the growing demand for higher capacity, speed, and reliability of memory products for AI workloads with its industrial-grade DRAM portfolio. ATP DRAM modules address memory bandwidth requirements in diverse deployment scenarios.
Innovative Applications and Future Prospects of AI in Antenna Design
2024-12-04 - Technical Discussion With the rapid development of technology, the complexity of telecommunications systems is increasing, and traditional design methods led by human engineers are facing unprecedented challenges. To address this trend, artificial intelligence (AI) is gradually occupying a central role in the design and operation of telecommunications systems.
AI Chip Shortage: Nvidia‘s Blackwell GPUs and the Future of AI Technology
2024-10-24 - Industry News Nvidia‘s Blackwell GPUs are sold out for 12 months due to high demand from major clients like Amazon, Dell, Google, Meta, and Microsoft. Memory chips like HBM are also in high demand, with manufacturers‘ 2024 production fully booked. TSMC‘s 3nm and 5nm process technologies are driving growth, but concerns about an AI market bubble and sustainability of Nvidia‘s stock growth exist. Despite this, strong demand for AI is expected to continue, driven by tech firms‘ significant investments in AI infrastructure.
ROHM Develops Ultra-Low-Power On-Device Learning Edge AI Chip，Enabling Real-time Failure Prediction Without Cloud Server Required
2022-12-01 - New Product Introduction ROHM has developed an on-device learning AI chip for edge computer endpoints in the IoT field. It utilizes artificial intelligence to predict failures (predictive failure detection) in electronic devices equipped with motors and sensors in real time with ultra-low power consumption.
Song Shiqiang from SLKOR Keynote Speech -How AI Empowers New Quality Productive Forces
2024-06-15 - Manufacturer News SLKOR and Kinghelm CEO, Mr. Song Shiqiang, interprets how AI Empowers New Quality Productive Forces - high-tech, high efficiency and high quality.
Generative AI and EV Batteries: Why Liquid Cooling?
2023-06-27 - Manufacturer News Advancing technologies like high-performance AI and EV batteries use more power. More power generates more waste heat, so much so that generative AI and EV battery innovators are shifting to liquid cooling. We‘ll explore why liquid cooling is a fundamental part of this conversation.
X-PHY® AI Embedded Cybersecure Solid-State Drive (SSD)
2024/12/20 - Supplier and Product Introduction
X-PHY® AI EMBEDDED CYBERSECURE SOLID-STATE DRIVE (SSD)
SMART ZDIMMs for AI and ML
10.06.24 - Supplier and Product Introduction
SRZHG8RD5648-SP,SRZ4G8RD5448-SP,SRZ2047RD410825-SE,SRZ8G8RD5448-SP,SRZ4G8RD5288-SP,SRZ4097RD420825-SC,SRZAG8RD5846-SB,SRZ8197RD440425-SC,SRZ4097RD440425-SC
查看更多版本
SMART NVDIMMs for AI and ML
10.19.21 - Datasheet
ST2047NJ420425HD,ST2047NJ420425HD1,ST2047NJP20425MJ2,STCA8062SMTH061B,ST2047NJP20425MJ1,STC24041SMCD011B,STCA8062SMTH062B,ST4097NJ44D472SB,STC48042SMTDE81BVG,STC54061INTH011B,ST2047NJ420414HD1,STC22041SMTD051B,ST2047NJP20425MJ,STC54061SMCH021B,STC36041INTH011B,STC48042SMTDE82BVG
查看更多版本
MWC 2025 | Fibocom Debuts “AI For X“: Reshaping Industries with AI
2025-06-03 - Manufacturer News As a global leading provider of AIoT modules and solutions, Fibocom has been committed to creating innovative IoT solutions for industries worldwide for the past 26 years. In response to the AI-driven era, Fibocom is introducing “AI For X,“ a strategy aimed at helping industries shift from merely “connecting everything” to “intelligently connecting everything.” This approach leverages a comprehensive suite of AI technologies, products, industry solutions, and ecosystem integration.
Zbit Participated SCGC 2023 and Began to Market the Best Cost-effective AI Products and Solutions
2023-09-12 - Manufacturer News Zhao xinlin, vice president of marketing of Zbit, attended the annual meeting as an invited guest. In the future, the company will combine its own technology accumulation in the field of NOR Flash and MCU, to develop integrated memory and Computing AI chip development and application.
Electronic Mall
Auth. Dist.
Auth. Dist.
Auth. Dist.
Auth. Dist.
Auth. Dist.
Auth. Dist.
Integrated Circuits
Discrete Components
Connectors & Structural Components
Assembly UnitModules & Accessories
Power Supplies & Power Modules
Electronic Materials
Instrumentation & Test Kit
Electrical Tools & Materials
Mechatronics
Processing & Customization
