Data Generation Device for Parallel Processing Padding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Parallel processing devices face complexity in handling data with dependencies, requiring reprocessing and padding, which necessitates intricate memory access and operation unit control for efficient data processing.
Innovation Solution
A data generation device that includes a controller and a data processing unit with multiplexers and registers to generate input data by padding original data with specific values or copying data, allowing for efficient output to parallel processing devices, thereby simplifying the processing operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If padding operation is performed in parallel processing device, then data dependency is handled, but device complexity increases
Solution Approach 1:
The data generation device performs padding operations in advance before the parallel processing device receives the data. By pre-processing the padding operation using multiplexers and registers, the complex memory access and operation unit control are shifted to the data generation device, allowing the parallel processing device to receive already-padded data without needing to handle data dependencies internally.
2Reliability
If data is padded with special values, then boundary conditions are handled, but processing efficiency decreases
Solution Approach 1:
The padding operation is performed in advance by the data generation device before data is sent to the parallel processing device. This preliminary action ensures that boundary conditions are already handled when data reaches the parallel processing device, eliminating the need for additional padding operations and special value handling during the main processing workflow, thus maintaining processing efficiency.
Solution Approach 2:
The data generation device acts as an intermediary between the data source and the parallel processing device. It receives original data, performs padding operations using multiplexers and registers, and outputs prepared data to the parallel processing device. This intermediary role handles the complexity of padding and boundary conditions separately, allowing the parallel processing device to focus on core processing operations without efficiency penalties.
3Reliability
If reprocessing of previously processed data is performed, then data dependency is satisfied, but time consumption increases
Solution Approach 1:
The data generation device performs all necessary padding and data preparation operations in advance before data is sent to the parallel processing device. By pre-processing data dependencies and preparing all required data in advance using multiplexers and registers, the system eliminates the need for reprocessing previously processed data during the main computational workflow, thus reducing time consumption while satisfying data dependency requirements.
Data Source
AI summary
Provided is a data generation device for generating input data to be inputted to a parallel processing device. The data generation device includes: a controller configured to output padding data; and a data processing device configured to receive original data and to generate the input data in which at least a portion of the original data is padded with the padding data. The data processing device includes: a first multiplexer configured to receive the padding data and the original data; a register configured to store data outputted from the first multiplexer; and a second multiplexer configured to receive data outputted from the first multiplexer and data stored in the register.