MIMD Architecture

Figure 9: Two MIMD Categories; Shared Memory and Message Passing MIMD Architectures.

• It was apparent that distributed memory is the only way efficiently to increase the number of processors managed by a parallel and distributed system.
• If scalability to larger and larger systems (as measured by the number of processors) was to continue, systems had to use distributed memory techniques.

• Two broad categories, see Figure 9:
  1. Shared memory. Processors exchange information through their central shared memory.
     • Because access to shared memory is balanced, these systems are also called SMP (symmetric multiprocessor) systems.
  2. Message passing. Also referred to as distributed memory. Processors exchange information through their interconnection network.
     • There is no global memory, so it is necessary to move data from one local memory to another by means of message passing.
     • This is typically done by a Send/Receive pair of commands, which must be written into the application software by a programmer
     • Data copying and dealing with consistency issues.

• Programming in the shared memory model was easier, and designing systems in the message passing model provided scalability.
• The distributed-shared memory (DSM) architecture began to appear in systems. In such systems,
  • memory is physically distributed; for example, the hardware architecture follows the message passing school of design,
  • but the programming model follows the shared memory school of thought.
  • Thus, the DSM machine is a hybrid that takes advantage of both design schools.