Processor Data Gathering via Pre-sorted Memory Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The processing time and efficiency decrease when gather or scatter instructions are used for vector operations due to the need for increased memory access, as only continuous memory areas can be accessed in one operation, leading to discontinuous memory access for data values.
Innovation Solution
A processor that loads or stores data elements by changing the storage order, allowing for efficient data gathering and storage by reading and writing data elements from adjacent memory areas collectively and sorting them into corresponding storage units, reducing the need for individual memory accesses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If gather or scatter instructions are used for vector operations, then data can be accessed from designated indices of a data array, but the number of memory accesses increases and processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-organizing data in the memory into continuous areas corresponding to vector operation units before the vector operation is executed. This preprocessing step groups data elements that will be accessed together during the gather operation, so that when the gather instruction is executed, all required data elements can be fetched in a single continuous memory access operation rather than requiring multiple separate accesses for discontinuous data locations.
2Adaptability or versatility
If gather or scatter instructions are used for vector operations, then data can be accessed from designated indices of a data array, but the number of memory accesses increases
Solution Approach 1:
The patent applies preliminary action by pre-organizing data in the memory into continuous areas corresponding to vector operation units before the vector operation is executed. This preprocessing step groups data elements that will be accessed together during the gather operation, so that when the gather instruction is executed, all required data elements can be fetched in a single continuous memory access operation rather than requiring multiple separate accesses for discontinuous data locations.
3Adaptability or versatility
If data is stored in discontinuous memory areas, then data values can be stored at designated indices, but memory access must be performed in respective clusters of continuous areas
Solution Approach 1:
The patent applies the inversion principle by reversing the conventional approach: instead of having data scattered in discontinuous memory areas and requiring complex clustered access patterns, the patent inverts the storage organization so that data is pre-arranged in continuous memory areas that correspond to the vector operation units. This inversion simplifies the memory access pattern from complex clustered access to simple sequential access, while still supporting the same data indexing functionality.
Data Source
AI summary
An object of the present invention is to efficiently perform a data load process or a data store process between a memory and a storage unit in a processor. The processor includes: a plurality of storage units associated with a plurality of data elements included in a data set; and a control unit that reads the plurality of data elements stored in adjacent storage areas from a memory, in which a plurality of the data sets is stored, collectively for respective data sets, sorts the respective read data elements to a storage unit corresponding to the data element among the plurality of storage units, and writes the data elements to the respective data sets.