Neural Network Forward Fusion Template Fuse Unit
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of layers and parameters in neural networks leads to significant on-chip and off-chip input/output accesses, consuming resources and delaying operation times, necessitating a mechanism to reduce these overheads.
Innovation Solution
An integrated circuit apparatus and method for forward fusion of a neural network, which creates a template fuse unit to perform neural network computing, reducing input/output overheads by fusing adjacent layers and optimizing data transfer between on-chip and off-chip memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of layers and parameters in neural network is increased, then the computing capability and accuracy are improved, but the on-chip and off-chip input/output accesses increase, consuming more resources and delaying operation time
Solution Approach 1:
The patent merges adjacent layers into a template fuse unit, combining multiple sequential operations into a single integrated computing unit. This reduces the number of separate input/output accesses between layers, thereby decreasing the time spent on data transfer while maintaining the computational capability of the original multi-layer structure.
2Reliability
If the number of layers and parameters in neural network is increased, then the computing capability and accuracy are improved, but the on-chip and off-chip input/output accesses increase, consuming more resources
Solution Approach 1:
By fusing adjacent layers into a template fuse unit, the patent reduces the frequency of data transfers between on-chip and off-chip memory. This consolidation decreases the total number of input/output operations, thereby reducing energy consumption associated with these transfers while preserving the computational functionality.
3Reliability
If multiple layers are used in neural network, then the feature extraction and classification performance are improved, but the data transfer overhead between on-chip and off-chip memory increases
Solution Approach 1:
The patent combines multiple adjacent layers into a template fuse unit that processes data in a single integrated operation. This reduces the number of intermediate data transfers between on-chip and off-chip memory, thereby reducing transfer overhead and energy consumption while maintaining the feature extraction capabilities through the fused layer structure.
Data Source
AI summary
The present disclosure relates to an apparatus and a method for forward fusing a neural network, a board card, and a readable storage medium. The computing apparatus of the present disclosure is included in an integrated circuit apparatus. The integrated circuit apparatus includes a general interconnection interface and other processing apparatus. The computing apparatus interacts with other processing apparatus to jointly complete a computing operation specified by a user. The integrated circuit apparatus further includes a storage apparatus. The storage apparatus is connected to the computing apparatus and other processing apparatus, respectively. The storage apparatus is used for data storage of the computing apparatus and other processing apparatus.