PCIe Switch Pre-Read Mechanism for Bandwidth Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The PCIe bus system experiences reduced bandwidth utilization due to increased data transmission delay and the limitation of 256 unresponsive read requests, leading to a cliff-like drop in bandwidth utilization as bandwidth and routing delay increase.
Innovation Solution
A data transmission method that converts Non-Posted data transmission mode to Posted data transmission mode by pre-reading data segments on a PCIe link, allowing multiple data segments to be provided in advance to a first processing device, which then returns the requested data segment as response data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the PCIe device ensures that the total number of Read TLPs on the PCIe bus is not more than 256, then data transmission stability is maintained, but bandwidth utilization drops cliff-like when bandwidth increases and routing delay increases
Solution Approach 1:
The patent introduces a pre-read mechanism where the PCIe switching device proactively reads data from the PCIe destination end before actual read requests arrive. By pre-reading multiple data segments and caching them locally, the system prepares data in advance, reducing the need for unresponsive read requests and improving bandwidth utilization while maintaining transmission stability
Solution Approach 2:
The patent introduces a PCIe switching device as an intermediary component between the PCIe request end and PCIe destination end. This intermediary performs pre-reading operations and maintains a cache of pre-read data, acting as a buffer that decouples the request rate from the response rate, thereby resolving the contradiction between stability limits and bandwidth utilization
2Reliability
If the PCIe device stops sending Read TLPs when the number exceeds 256 and waits for Completion TLP, then data stability is ensured, but data transmission delay increases
Solution Approach 1:
The PCIe switching device performs pre-reading of data segments before actual read requests are processed. By having data ready in the cache in advance, the system eliminates waiting time for Completion TLPs while maintaining data stability through controlled pre-reading operations
Solution Approach 2:
The patent enables continuous data transmission by maintaining a pipeline of pre-read operations. While some Read TLPs are being processed, other pre-read data is already available in the cache, ensuring continuous useful action without interruption and reducing overall transmission delay
Data Source
AI summary
A data transmission method, a device, a network system, and a storage medium. According to the method, a PCIe request end requests an application scenario of a data fragment from a PCIe destination end according to a Non-Posted data transmission mode; and on a PCIe link, the Non-Posted data transmission mode is converted, without the perception of the PCIe request end and PCIe destination end, into a Posted data transmission mode by means of pre-reading, so that the delay of the PCIe link is reduced, and the utilization of bandwidth of the PCIe link is improved.