XOP(Cross Operating Processor)
The Cross Operating Processor (XOP) is a processor specialized in data comparison.
XOP’s concept
The XOP is a processor specialized in data comparison. When two sets of data become large, a CPU has to execute extremely severe operations, but the XOP can compare two sets of XY data directly, which is much more efficient than CPU processes.
XOP’s chip configuration
The Xd and Yd in the diagram are the memory (data) in the line direction and column direction, respectively.
This processor has operators that compare the data at the intersection of the matrix data described above.
If there are 1K pieces of data in both line X and column Y, the number of operators is 1M.
Since 1M operators perform SIMD-type comparison operations, matching and similar data are found extremely fast.
If either Ｘ or Ｙ is treated as known data (reference), it can be used in a histogram or classification (grouping).
If both Ｘ and Ｙ are treated as unknown data, it is possible to perform clustering processes such as name identification without an index.
XOP application examples
XOP can not only compare two types of unknown data through a histogram, sorting, correlation, and classification, but also completely eliminate the burden of information processing applied to the CPU by treating either X or Y as known data.
