-
Book Overview & Buying
-
Table Of Contents
Modern Computer Architecture and Organization - Third Edition
By :
Some algorithms benefit from explicitly vectorized execution, which means the same operation is performed simultaneously on multiple data items. Processor instructions tailored to such tasks are called SIMD instructions. SIMD execution contrasts with the behavior of superscalar processors, which typically issue multiple instructions at once, each performing a different operation on different data—an example of multiple-instruction, multiple-data (MIMD) parallel processing. Some common algorithms, such as the dot product used in digital signal processing (described in Chapter 6, Specialized Computing Domains), perform the same mathematical operation on an array of values and benefit significantly from SIMD processing.
While the multiply-accumulate (MAC) operation described in Chapter 6 performs scalar mathematical operations on each pair of vector elements in sequence, it is also possible to implement processor hardware and instructions capable of performing the...