Scribd
Upload
39 views22 pages

Outline: - Course Administration 18.337/6.338/SMA5505 - Parallel Machines in 2003

This document provides an overview of parallel computer architectures and applications. It discusses different types of parallel machines including massively parallel processors (MPPs), clusters, symmetric multiprocessors (SMPs), and constellations. It notes that clusters are becoming more prominent as they are cheaper to build than custom MPPs. The document also outlines trends in parallel computing performance and applications, and mentions approaches like distributed computing and embedded systems.

Uploaded by

Web dev
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
39 views22 pages

Outline: - Course Administration 18.337/6.338/SMA5505 - Parallel Machines in 2003

This document provides an overview of parallel computer architectures and applications. It discusses different types of parallel machines including massively parallel processors (MPPs), clusters, symmetric multiprocessors (SMPs), and constellations. It notes that clusters are becoming more prominent as they are cheaper to build than custom MPPs. The document also outlines trends in parallel computing performance and applications, and mentions approaches like distributed computing and embedded systems.

Uploaded by

Web dev
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 22

Outline

• Course Administration 18.337/6.338/SMA5505

• Parallel Machines in 2003


– Overview
– Details
• Applications
• Special Approaches
• Our Class Computer
Parallel Computer Architectures
• MPP – Massively Parallel Processors
– Top of the top500 list consists of mostly mpps but
clusters are “rising”

• Clusters are there!


– Earth Simulator: Old-old style making news again
– ASCI Machines: Big companies, special purpose
– Beowulf Clusters: Popping up everywhere
• Software
– Embarrassingly parallel or sacrifice a grad student
– MATLAB*p (our little homegrown project)
Always interesting to check out http://www.top500.org/dlist/2002/11/
Performance Trends

Courtesy of Top500.org.
Extrapolations

Courtesy of Top500.org.
•Notes by Jack Dongarra
Earth Simulator
Beowulf Clusters
Current Beowulfs (2)
Parallel Computer Architectures
• MPP – Massively Parallel Processors
– Top of the top500 list consists of mostly mpps but
clusters are “rising”
• Clusters
– “simple” cluster (1 processor in each node)
– Cluster of small smp’s (small # processors / node)
– Constellations (large # processors / node)
• Older Architectures
– SIMD – Single Instruction Multiple Data (CM2)
– Vector Processors (Old Cray machines)
Architecture Details
1. MPPs are built with specialized networks by
vendors with the intent of being used as a
parallel computer. Clusters are built from
independent computers integrated through an
aftermarket network.
• Buzzwords: “COTS” Commodity off the shelf
components rather than custom architectures.
• Clusters are a market reactions to MPPS with the
thought of being cheaper
• Originally considered to have slower
communications but are catching up.
More details
• 2. “NOW”-Networks of Workstations
• Beowulf (Goddard in Greenbelt MD) –
• Clusters of a small number of pc’s
• (pre 10th century poem in Old English about
a Scandinavian warrior from the 6th
century)
The First Beowulf (2)
Architecture Details
• Computers Æ MPPs
P M World’s simplest computer (processor/memory)

C
P M D Standard computer (add cache,disk)

C
P M D
Network
C
P M D
C
P M D
Architecture Details SMP
(Symmetric Multiprocessor)
P/C P/C

M + Disk

P/C P/C

NUMA – Non uniform memory access


4. Constellation: Every node is a large smp
More details
• 5) SIMD – SIngle instruction multiple data
• data parallel vs mimd
• 6: Speeds:

• Megaflops 106 flops


• Gigaflops 109 flops workstations
• Teraflops 1012 top 17 supercomputers
by 2005 every supercomputer in the top 500
• Petaflops 1015 2010?
Trends
• Moore’s Law: Number of Transistors per
square inch in an integrated circuit doubles
every 18 months
• Every decade – computer performance
increases 2 order of magnitude
Applications of Parallel
Computers
• Traditionally: government labs, numerically
intensive applications
• Research Institutions
• Recent Growth in Industrial Applications
– 236 of the top 500
– Financial analysis, drug design and analysis, oil
exploration, aerospace and automotive
Goal of Parallel Computing
• Solve bigger problems faster
• Often bigger is more important than faster
• P-fold speedups not as important!

Challenge of Parallel Computing


Coordinate, control, and monitor the computation
Easiest Applications
• Embarrassingly Parallel – Lots of work that
can be dived out with little coordination or
communication
• Example: integration, Monte Carlo
methods, Adding numbers
Special Approaches
• Distributed Computing on the internet
– signal processing 15 Teraflops
– Distributed.net factor product of 2 large primes
– Parabon – biomedical, protein folding, gene expression
• Akamai Network – Tom Leighton, the late Danny
Lewin
• Thousands of servers spread globally that caches
web pages and routes traffic away from congested
areas
• Embedded Computing : Mercury (inverse to the
worldwide distribution)
The Computational Grid
• Computers are everywhere. Can we plug into the
power? Challenges: managing machines that are
distributed, might fail, be turned off, or there
might be malicious intent.
MATLAB*P demo

You might also like