Hardware Arbitration Unit for Semiconductor CPU Load Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing arbitration methods for semiconductor devices require software processing, leading to increased CPU load and response time due to the need for communication between CPUs and operating systems, which becomes inefficient as hardware IP speeds increase.
Innovation Solution
A semiconductor device with a hardware IP that includes control receiving units and an arbitration unit to manage operation requests from software modules, allowing the hardware IP to handle arbitration independently, reducing the CPU's load by controlling information transmission and processing only one operation request at a time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If arbitration processing is performed by software communication between CPUs and operating systems, then the system can manage shared resources, but the CPU load increases and response performance deteriorates
Solution Approach 1:
The arbitration function is extracted from the software domain and implemented as a dedicated hardware arbitration unit. This unit receives operation requests from multiple control receiving units and autonomously selects which request to forward to the calculation unit, eliminating the need for CPU intervention in arbitration decisions and thereby reducing CPU load while improving response performance
Solution Approach 2:
A hardware arbitration unit is introduced as an intermediary component between the control receiving units and the calculation unit. This intermediary handles the arbitration of operation requests in hardware, preventing direct CPU involvement in the arbitration process and thus resolving the contradiction between reliable shared resource management and maintained CPU throughput
2Speed
If hardware IP speed increases, then processing capability improves, but software arbitration response frequency must increase leading to higher CPU load
Solution Approach 1:
The software-based arbitration mechanism is replaced with a hardware-based arbitration unit that operates independently of CPU cycles. This hardware arbitration unit can respond to operation requests at the same high speed as the hardware IP, eliminating the bottleneck where software arbitration response frequency had to increase with hardware IP speed, thus avoiding increased CPU load and arbitration processing complexity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In a semiconductor device, a load of CPU required for arbitration when using a shared resource is reduced. The semiconductor device includes a CPU section and a hardware IP. In the CPU section, software modules are executed. The hardware IP includes a storage unit, an arbitration unit, and a calculation unit. The storage unit includes control receiving units that receive operation requests transmitted by the software modules, respectively. The calculation unit performs processing based on an operation request transmitted from the control receiving units. The arbitration unit controls information transmission between the control receiving units and the calculation unit so that the calculation unit receives only an operation request from any one of the control receiving units.