Integrated Multiply-Accumulate Module for Fixed-Point and Floating-Point Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing chip designs require separate hardware structures for fixed-point and floating-point arithmetic modes, leading to increased chip area and power consumption due to the need for multiple independent multiply accumulate modules, which complicates manufacturability, yield, and performance.
Innovation Solution
A chip with a single multiply accumulate module that integrates a fixed-point general-purpose unit and a floating-point special-purpose unit, allowing for both fixed-point and floating-point arithmetic modes, sharing some devices to reduce area and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate hardware structures are used for fixed-point and floating-point arithmetic modes, then both operation modes can be supported, but chip area and power consumption increase
Solution Approach 1:
The patent merges fixed-point and floating-point multiply-accumulate modules into a single integrated structure. The fixed-point MAC module and floating-point MAC module share common components including multiplication units, addition units, and control logic, allowing both arithmetic modes to be supported while reducing overall chip area and power consumption compared to having completely separate hardware structures
Solution Approach 2:
The integrated MAC module is designed with multi-functionality to perform both fixed-point and floating-point operations. By incorporating mode selection logic and shared computational resources, the single module can adaptively execute different arithmetic operations based on the input mode signal, eliminating the need for dedicated separate hardware for each operation type
2Adaptability or versatility
If separate hardware structures are used for fixed-point and floating-point arithmetic modes, then both operation modes can be supported, but power consumption increases
Solution Approach 1:
The patent merges fixed-point and floating-point multiply-accumulate modules into a single integrated structure. The fixed-point MAC module and floating-point MAC module share common components including multiplication units, addition units, and control logic, allowing both arithmetic modes to be supported while reducing overall chip area and power consumption compared to having completely separate hardware structures
Solution Approach 2:
The integrated MAC module is designed with multi-functionality to perform both fixed-point and floating-point operations. By incorporating mode selection logic and shared computational resources, the single module can adaptively execute different arithmetic operations based on the input mode signal, eliminating the need for dedicated separate hardware for each operation type
3Adaptability or versatility
If multiple independent multiply accumulate modules are used, then both fixed-point and floating-point operations can be performed, but device complexity increases
Solution Approach 1:
The patent merges fixed-point and floating-point multiply-accumulate modules into a single integrated structure. The fixed-point MAC module and floating-point MAC module share common components including multiplication units, addition units, and control logic, allowing both arithmetic modes to be supported while reducing overall chip area and power consumption compared to having completely separate hardware structures
Data Source
AI summary
A chip includes a multiply accumulate module with a fixed-point general-purpose unit, a floating-point special-purpose unit, and an output selection unit. The fixed-point general-purpose unit and the floating-point special-purpose unit share one group of multipliers. In the multiply accumulate module of the chip, the fixed-point operation and the floating-point operation are integrated in one circuit, so that the multiply accumulate module implements not only the fixed-point operation, but also the floating-point operation in the circuit. Sharing the multiplier by the fixed-point operation unit and the floating-point operation unit reduces a total quantity of devices used, and reduces power consumption during operation.


