Parallelism

Parallelism

• 1.
  Welcome to ourpresentation Parallelism
• 2.
  Parallelism Goals of Parallelism Exploitationof Concurrency Types of Parallelism Md. Monirul Awal (161-15-7501) Md. Raseduzzaman (161-15-7495)
• 3.
  Parallelism Executing two ormore operations at the same time is known as Parallelism.
• 4.
  Goals of Parallelism The purpose of parallel processing is to speedup the computer processing capability or in words, it increases the computational speed.  Increases throughput, i.e. amount of processing that can be accomplished during a given interval of time.  Improves the performance of the computer for a given clock speed.  Two or more ALUs in CPU can work concurrently to increase throughput.  The system may have two or more processors operating concurrently.
• 5.
  Exploitation of Concurrency Techniquesof Concurrency:  Overlap : execution of multiple operations by heterogeneous functional units.  Parallelism : execution of multiple operations by homogenous functional units. Throughput Enhancement:  Internal Micro-operations: performed inside the hardware functional units such as the processor, memory, I/O etc.  Transfer of information: between different functional hardware units for Instruction fetch, operand fetch, I/O operation etc.
• 6.
  Types of Parallelism: 1.Instruction Level Parallelism (ILP) 2. Processor Level Parallelism
• 7.
  1. Instruction LevelParallelism (ILP) Instruction-level parallelism (ILP) is a measure of how many operations in a computer program can be performed "in-parallel" at the same time.
• 8.
  ILP TECHNIQUES An instructionpipeline is a technique used in the design of modern microprocessors, microcontrollers and CPUs to increase their instruction throughput (the number of instructions that can be executed in a unit of time). For example, the RISC pipeline is broken into five stages with a set of flip flops between each stage as follow:
• 9.
  ILP TECHNIQUES A superscalarCPU architecture implements ILP inside a single processor which allows faster CPU throughput at the same clock rate Simple superscalar pipeline. By fetching and dispatching two instructions at a time, a maximum of two instructions per cycle can be completed.
• 10.
  2. Processor LevelParallelism  Instruction-level parallelism (pipelining and superscalar operation) rarely win more than a factor of five or ten in processor speed.  To get gains of 50, 100, or more, the only way is to design computers with multiple CPUS  We will consider three alternative architectures: –Array Computers –Multi-processors
• 11.
  Array Computer  Anarray processor consists of a large number of identical processors that perform the same sequence of instructions on different sets of data.  A vector processor is efficient at executing a sequence of operations on pairs of Data elements; all of the addition in a single, heavily-pipelined adder.
• 12.
  Multi-processor The processing elementsin an array processor are not independent CPUS, since there is only one control unit.  The first parallel system with multiple full-blown CPUs is the multiprocessor.  This is a system with more than one CPU sharing a common memory coordinated in software.  The simplest one is to have a single bus with multiple CPUs and one memory all plugged into it.