Mumbai University > Computer Engineering > Sem 8 > parallel and distributed systems

Processing element (PE) in SIMD machines-

1. General register: There are 'n' no. of registers from R1 to Rn storing values for processing.
2. ALU: It is a Arithmetic Logic Unit for computation.
3. Mask: It is a mask bit the result of a logical operation between vectors for PE as required by algorithm.
4. Index Register: The content of this register could be added to the global address provided by CU in the instruction.
5. Address Register: The address of system memory form where the PE access the data.
6. Communication Section: It provides the communication between PE and other PEs and global memory system.
7. Local memory: It is provided for temporary storage of data.

Each PE receives the instruction broadcast by CU. Decoder unit and execution unit if it is not marked. If the PE is marked then it executes the instruction as NOP.

Symbol for PE in Mesh connected network-