Chapter 2: CS621 1
Parallel and Distributed Computing
Chapter 2: Parallel Programming Platforms
Jun Zhang
Laboratory for High Performance Computing & Computer Simulation
Department of Computer Science
University of Kentucky
Lexington, KY 40506
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Parallel Programming Platforms of Parallel and Distributed Computing | CS 621, Study notes of Computer Science
Material Type: Notes; Professor: Zhang; Class: PARALLEL AND DISTRIBUTED COMPUTING; Subject: Computer Science; University: University of Kentucky; Term: Unknown 1989;
Typology: Study notes
Pre 2010
1 / 65
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download Parallel Programming Platforms of Parallel and Distributed Computing | CS 621 and more Study notes Computer Science in PDF only on Docsity!
Chapter 2: CS
Parallel and Distributed ComputingChapter 2: Parallel Programming Platforms
Jun Zhang
Laboratory for High Performance Computing & Computer Simulation
Department of Computer Science
University of Kentucky
Lexington, KY 40506
Chapter 2: CS 2.1a: Flynn’s Classical Taxonomy
One of the more widely used parallel computerclassifications, since 1966, is called Flynn’s TaxonomyIt distinguishes multiprocessor computers according tothe dimensions of
Instruction
and
Data
SISD
: Single instruction stream, Single data stream
SIMD
: Single instruction stream, Multiple data streams
MISD
: Multiple instruction streams, Single data stream
MIMD
: Multiple instruction streams, Multiple data
streams
Chapter 2: CS
2.2a: SIMD Machines (I)
A type of parallel computers
Single instruction: All processor units execute the sameinstruction at any give clock cycle
Multiple data: Each processing unit can operate on a differentdata element
It typically has an instruction dispatcher, a very high-bandwidthinternal network, and a very large array of very small-capacityinstruction units
Best suitable for specialized problems characterized by a highdegree of regularity, e.g., image processing
Two varieties: Processor Arrays and Vector Pipelines
Examples: Connection Machines, MasPar-1, MasPar-2;
IBM 9000, Cray C90, Fujitsu VP, etc
Chapter 2: CS
2.2b: SIMD Machines (II)
Chapter 2: CS
2.2d: Processing Array
Chapter 2: CS
2.2e: MasPar Machine
Chapter 2: CS
2.2g: Assembly Line
Chapter 2: CS
2.2h: Vector Processor Pipeline
Chapter 2: CS 2.3b: MISD Machines (II)
Chapter 2: CS
2.4a: MIMD Machines (I)
Multiple instruction: Every processor mayexecute a different instruction stream
Multiple data: Every processor may work witha different data stream
Execution can be synchronous orasynchronous, deterministic or non-deterministic
Examples: most current supercomputers,grids, networked parallel computers,multiprocessor SMP computer
Chapter 2: CS
2.4c: MIMD Machines (III)
Chapter 2: CS 2.4d: MIMD Machines (T3E-Cray)
Chapter 2: CS
2.6a: Shared Memory Computers
All processors have access to all memory asa global address space
Multiple processors can operateindependently, but share the same memoryresources
Changes in a memory location effected byone processor are visible to all otherprocessors
Two classes of shared memory machines:UMA and NUMA, (and COMA)
Chapter 2: CS