The evolution of supercomputers has marked significant advancements from the Cray-1, a single-CPU machine, to modern parallel computers with thousands of CPUs. These powerful machines can perform trillions of floating-point operations per second and have become more cost-effective and accessible due to the use of commodity off-the-shelf components, exemplified by the Beowulf system.
The first of these supercomputers, ASCI Red, was delivered to Sandia National
Laboratories in 1997
Beowulf
Beowulf is all example of a system wnstlUcted out of commodity, off-theshelf
(COTS) componenls.
Their system, named Beowulf, contained
16 Intel DX4 processors connected by multiple to Mbitfsec Ethernet links
The dynamic PC marketplace set the
stage for the next breakthrough in parallel computing
Commercial Parallel_Computers
Some commercial parallel computers had support for higher-level parallel
programming languages and debuggers
Experimenal parallel computers were less expensive
The Cosmic Cube
Intel Corporation had donated much of the hardware for the Cosmic Cube.
the potential for microprocessor-based parallel computing
, a parallel computer constructed out of 64 Intel 8086 microprocessors
[34]
SEEKING CONCURRENCY
Functional Parallelism
a<-2
17 <- 3
m <- (a +b)/2
s <- (a2 + 172
)/2
v<- s ~ m2
Data Parallelism
for i <- 0 to 99 do
a[i] <- b[i] + e[i]
endfor
DATA CLUSTERING
Multidimensional data clustering is an important tool for data mining.
make info
easier to remember
interesting
save time
Subtema
Data Clostering:
1. OUlPUT K centers
1. Repeat the following steps for I iteratiolis or until he performance function converges. whichver
come firts
3. Choose the K initial cluster cenlers using II rnndom sample
yellow = facts, examples
2. For each of the N documenLI .generate a D-dirnensional vector indicating how well il cover the
D different topics
1. input N documents
PROGRAMM1NG PARALLEL COMPUTERS
Create a Parallel Language
Add a Parallel Programming Layer
Extend a Sequential Programming Language
Extend a Compiler
EVOLUTION OF SUPERCOMPUTING
Supercomputers have typically cost $10 million or more.
Even mainframe computers
are being constructed out of microprocessors.
Today, supercomput£r means a parallel computer with thousand~ of CPUs.
Supercomputers are the most powerful computers that can be built.
In 1976
supercomputer meant a Cray-l, a single-CPU computer with a high ·petformance
pipelined Vector processor wnnected to a lIigh-performance memory system.
Today's supercomputers
are more than a billion times faster, able to petfonn trillions of floating point
operations per second
MODERN SCIENTIFIC METHOD
Global weather and environmental modeling
Medicine, and modeling of human organs and bones
Biology, pharmacology, genome sequencing, genetic engineering, protein
folding, enzyme activity, and cell modeling
Materials design and supercond.uctivity
Computational fluid dynamics and turbulence
Cosmology and astropbysics
you'll come up with original ideas
Quantum chemistry, statistical mechanics, and relativistic physics
Explain the phenomenon and designs an expellment to test that theory
Whats parallel computing?
Why should I program using MPI and OpenMP?
MPI is a perfectly satisfactory way for processors in different SMPs to cornrnullicate
with each other, OpenMP is a better way for processors within a single SMP to interact
MPI (Message Passing Interface) is a standard specification for messagepassing
lihmries
Reduce the time needed
to solve a single computational problem.
is a multiple-processor computer system suppmting
parallel programming.