Skip to main content

Types of array processing

Short answer

Attached array processor is an auxiliary processor connected via I/O, performing array computations with its own local memory. SIMD array processor has multiple processing elements controlled by one unit, all executing the same instruction on different data.

Describe the different types of array processing.

An array processor is a processor that performs computations on large arrays of data. The term is used for two different types of processor.

1. Attached array processor

An auxiliary processor attached to a general-purpose computer to improve its performance in numerical work. It connects to the host through an I/O interface and has its own local memory, which is linked to the main memory by a high-speed bus. The host computer does the general work, while the attached processor does array and matrix computations (for example, FFT and matrix multiplication) using multiple pipelined functional units.

 General-purpose   I/O interface   Attached array
 computer  <--------------------> processor
     |                                 |
 Main memory <---- high speed memory bus ---> local memory

2. SIMD array processor

A computer with many identical processing elements (PEs) under one control unit. Each PE has a local memory. The master control unit sends the same instruction to all PEs, and each works on its own data (Single Instruction, Multiple Data). Masking schemes can disable some PEs. An example is the ILLIAC IV with 64 PEs.

        Master control unit
        /     |      |     \
     PE1     PE2    PE3 ... PEn
      |       |      |       |
     M1      M2     M3      Mn

SIMD array processors are good for vector and matrix operations; attached array processors improve an existing machine at a low cost.

✓ Answer checked by Claude Opus 5.5 (high effort)

All Computer Organization and Architecture chapter 4 questions →