Fixed-Point ML Computing Architecture for Faster Parallel Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
General purpose processors are inefficient in processing large loads, leading to delays in tasks such as machine learning, particularly in training artificial intelligence models, due to their limited operating speed.
Innovation Solution
A computation device and method that includes a controller unit, operation unit, and conversion unit, utilizing a primary processing circuit and secondary processing circuits to perform machine learning computations by converting input data into fixed-point data and executing operations to improve processing speed and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If general purpose processors are used to process information, then the device complexity is low and ease of operation is maintained, but the processing speed and efficiency deteriorate under large loads
Solution Approach 1:
The processor is divided into a control unit that manages instructions and multiple operation units that perform parallel computations. This segmentation allows the system to maintain simple individual components while achieving high overall processing throughput through coordinated parallel operation.
Solution Approach 2:
The patent transitions from sequential single-threaded processing to parallel multi-threaded processing by adding multiple operation units. This dimensional change from one-dimensional sequential execution to multi-dimensional parallel execution dramatically increases processing capacity without proportionally increasing control complexity.
2Productivity
If general purpose processors operate at high speed, then processing efficiency improves, but power consumption and heat generation worsen
Solution Approach 1:
Different operation units are assigned specialized functions optimized for specific computational tasks. Each unit operates at optimal speed and power consumption for its designated function, avoiding the energy waste of general-purpose high-speed operation across all units. The control unit coordinates these specialized units to achieve high overall efficiency with reduced total power consumption.
Data Source
AI summary
The present disclosure provides a computation device. The computation device is configured to perform a machine learning computation, and includes an operation unit, a controller unit, and a conversion unit. The storage unit is configured to obtain input data and a computation instruction. The controller unit is configured to extract and parse the computation instruction from the storage unit to obtain one or more operation instructions, and to send the one or more operation instructions and the input data to the operation unit. The operation unit is configured to perform operations on the input data according to one or more operation instructions to obtain a computation result of the computation instruction. In the examples of the present disclosure, the input data involved in machine learning computations is represented by fixed-point data, thereby improving the processing speed and efficiency of training operations.