Neural Processing Unit Integer Multiplication Circuit
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic devices equipped with artificial intelligence technology face inefficiencies in performing operations related to artificial intelligence, particularly in handling different data types and performing multiplication operations between integer numbers of varying types.
Innovation Solution
An operational circuit within the electronic device is designed to combine and process integer numbers of different data types, utilizing a combiner, multiplier, and converter to efficiently perform multiplication operations, enabling the device to handle various data types such as floating-point numbers and integers, and accumulate results for neural network operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the operational circuit is designed to handle multiple data types (floating-point and integer), then the adaptability of the device is improved, but the device complexity increases
Solution Approach 1:
The operational circuit is designed with universal components that can handle multiple data types. The combiner, multiplier, and converter are configured to process both floating-point numbers and integer numbers through the same hardware pathway, allowing a single circuit to perform multiple functions rather than requiring separate dedicated circuits for each data type.
Solution Approach 2:
The operational circuit employs dynamic configuration where the combiner, multiplier, and converter can adapt their operation based on the input data type. The circuit dynamically adjusts its processing mode to handle either floating-point or integer operations, enabling flexibility without requiring static dedicated hardware for each data type.
2Productivity
If the combiner combines multiple integer numbers for multiplication, then the productivity of AI operations is improved, but the manufacturing precision of the operation results may deteriorate
Solution Approach 1:
The converter acts as an intermediary component between the multiplier and the final output. It receives the product from the multiplier and performs type conversion to generate the appropriate data type (floating-point or integer) for the result. This intermediary ensures that the combination of multiple integer multiplications maintains precision by properly handling the conversion process.
3Adaptability or versatility
If the converter outputs multiple data types from the multiplier, then the adaptability is improved, but the difficulty of detecting and measuring the output type increases
Solution Approach 1:
The system performs preliminary determination of the required output data type before the conversion process. The converter is configured based on the expected result type (floating-point or integer) before executing the multiplication operation, so the output type is predetermined and easily identifiable rather than requiring complex detection after the fact.
Data Source
AI summary
An operational circuit may include a combiner to combine, based on a request for multiplication of integer numbers different from floating-point numbers, a first integer number and a second integer number. The operational circuit may include a multiplier including first and second ports. A third integer number may be inputted to the first port, and a fourth integer number indicating a combination of the first integer number and the second integer number may inputted to the second port. The operational circuit may include a converter to output, based on a fifth integer number indicating a multiplication of the third integer number and the fourth integer number from a third port of the multiplier, a sixth integer number indicating a multiplication of the first integer number and the third integer number, and a seventh integer number indicating a multiplication of the second integer number and the third integer number.