Star Interconnection Architecture for SMP Cooling and Trace Length
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing symmetric multiprocessor (SMP) systems face challenges in interconnection design, as traditional backplane connectors are costly and create cooling difficulties due to crisscross airflow paths when CPUs are arranged orthogonally, limiting the number of CPUs that can be integrated while maintaining electrical performance.
Innovation Solution
A star interconnection architecture with a middle plane and orthogonally configured processor boards, utilizing traditional backplane connectors and a crisscross airflow cooling system to manage airflows effectively, allowing for higher CPU density without excessive trace lengths or cooling issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional backplane connectors are used to connect CPUs on a backplane, then the system cost is reduced, but the trace length between CPUs increases beyond maximum allowable limits
Solution Approach 1:
The patent transitions from a two-dimensional backplane layout to a three-dimensional orthogonal architecture. Processor boards are arranged in multiple layers with vertical interconnections through the middle plane, allowing traditional connectors to be used while maintaining short trace lengths through the vertical dimension.
Solution Approach 2:
The patent implements a nested structure where processor boards are positioned on opposite surfaces of a middle plane, with interconnects passing through the middle plane to connect corresponding boards. This nested arrangement enables short trace lengths while supporting multiple CPUs using traditional connectors.
2Reliability
If CPUs are arranged in an orthogonal configuration to reduce trace length, then electrical performance is improved, but cooling becomes difficult due to crisscross airflow paths
Solution Approach 1:
The patent divides the cooling system into separate modules positioned at different locations. First cooling modules are positioned to cool processor boards on one surface, while second cooling modules cool boards on the opposite surface, allowing independent cooling paths that avoid crisscross airflow interference.
Solution Approach 2:
The patent implements localized cooling by positioning cooling modules in proximity to specific processor boards they serve. Each cooling module creates dedicated airflow paths for its associated boards, ensuring effective cooling without interfering with other cooling zones in the orthogonal architecture.
3Reliability
If CPUs are placed very close together on a backplane to meet maximum trace length requirements, then electrical performance is maintained, but the number of CPUs that can be integrated is limited
Solution Approach 1:
The patent utilizes the third dimension by stacking processor boards on opposite surfaces of a middle plane with vertical interconnections. This three-dimensional arrangement allows integration of more CPUs without increasing trace lengths, as connections pass through the middle plane rather than traveling across the backplane surface.
Solution Approach 2:
The patent segments the CPU integration into multiple groups on opposite surfaces of the middle plane. Each surface can accommodate multiple processor boards independently connected through the middle plane, effectively multiplying the CPU capacity while maintaining short trace lengths for each connection.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the use of lower-cost connectors in high-performance configurations while maintaining efficient airflow and shorter trace lengths, supporting a higher number of CPUs within the system.
Implementation Method 1
The cooling system includes a first cooling module and a second cooling system module configured for generating a plurality of first airflows and second airflows for the first processor boards and the second processor boards respectively
Data Source
AI summary
A symmetric multiprocessor computer is provided with a star interconnection architecture and a cooling system. The star interconnection architecture include a middle plane, and plural first processor boards and second processor boards configured vertically onto opposite surfaces of the middle plane. The first processor boards and the second processor boards are crisscross to each other at the opposite surfaces of the middle plane. The cooling system includes a first cooling module and a second cooling system module configured for generating a plurality of first airflows and second airflows for the first processor boards and the second processor boards respectively, wherein the paths of the first airflows and the second airflows are crisscross to each other at the opposite surfaces of the middle plane.