Spatial Processor Buffering for Out-of-Order Stalled Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In spatial architectures, data requested by processing elements can be returned out of order, causing data to become stuck and leading to stalls that can halt the entire system.
Innovation Solution
Implement auxiliary memory control circuitry to monitor the network for stalled data and transfer it to an auxiliary storage buffer dynamically selected from local storage sectors based on usage metrics, reducing the likelihood of network stalls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data transfer requests are issued to the network in spatial architecture, then data can be retrieved from the network, but out-of-order data returns cause data to stall and potentially halt the entire system
Solution Approach 1:
The local storage is divided into multiple sectors, allowing the system to segment the storage space and dynamically allocate specific sectors for handling stalled data. This segmentation enables isolated management of different data types and prevents stalls from affecting the entire storage system.
Solution Approach 2:
The system dynamically selects which local storage sector to use as the auxiliary storage buffer based on current usage metrics. This dynamic allocation allows the system to adapt to varying data patterns and stall conditions, optimizing performance while maintaining stability.
2Reliability
If auxiliary storage buffer is dynamically selected from local storage sectors, then stalled data can be managed effectively, but storage allocation complexity increases
Solution Approach 1:
The memory control circuitry automatically monitors local storage usage metrics and autonomously selects the appropriate sector for the auxiliary storage buffer without external intervention. This self-service mechanism simplifies the overall system architecture by eliminating the need for complex manual configuration or external control.
Solution Approach 2:
The system changes the allocation parameter by selecting different local storage sectors based on usage metrics. This parameter-based approach allows flexible adaptation to different data access patterns while maintaining a relatively simple and manageable storage allocation mechanism.
Data Source
AI summary
There is provided a data processing apparatus and method. The data processing apparatus comprises a plurality of processing elements connected via a network arranged on a single chip to form a spatial architecture. Each processing element comprising processing circuitry to perform processing operations and memory control circuitry to perform data transfer operations and to issue data transfer requests for requested data to the network. The memory control circuitry is configured to monitor the network to retrieve the requested data from the network. Each processing element is further provided with local storage circuitry comprising a plurality of local storage sectors to store data associated with the processing operations, and auxiliary memory control circuitry to monitor the network to detect stalled data (S60). The auxiliary memory control circuitry is configured to transfer the stalled data from the network to an auxiliary storage buffer (S66) dynamically selected from amongst the plurality of local storage sectors (S64).