1. MICS ASSIGNMENT (B)
HPC
SUBMITTED BY :
PRIYANKA SINHA (12609210)
ARPIT GOYAL (12609159)
28/11/2012 HPC
1

2. HPC
NOTJUSTFOR ROCKET SCIENTIST
• Earlier the word supercomputer to someone was monstrously
complicated machines solving problems no one really understands.
• supercomputing was for the serious scientists and engineers who
needed to crunch numbers as fast as possible.
Today we have a different world. The custom supercomputer of
yesteryear has given way to commodity-based supercomputing , or
what is now called High Performance Computing (HPC).
28/11/2012
HPC
2

3. Featuresof HPC
• HPC is now considered as an essential part of business success.
• Can use the same hardware found in Web servers and even desktop
workstations.
• Cost effective
• Provide a great competitive advantage by designing new products,
optimizing manufacturing and delivery processes, solving production
problems, mining data , and simulating everything from business
process to shipping crates all in an effort to become more
competitive, profitable , and “green.”
28/11/2012
HPC
3

4. TheFormsof HPC
✓ The commodity HPC cluster
✓ Dedicated supercomputer
✓ HPC cloud computing
28/11/2012 HPC
HPC
4

5. WhoUsesHPC Today?
• The worldwide HPC market is growing rapidly. According to IDC, the total
HPC market was $10 billion in 2007 and is expected to hit $11.7 billion
around 2012.
The following is a list of major areas where HPC has a significant
presence:
 Bio-sciences and the human genome: Drug discovery, disease
detection/prevention.
Computer aided engineering (CAE): Automotive design and testing,
transportation, structural, mechanical design
 Chemical engineering: Process and molecular design
 Digital content creation (DCC) and distribution: Computer aided
graphics in film and media
28/11/2012
HPC
5

6.  Economics/financial: Wall Street risk analysis, portfolio management,
automated trading
 Electronic design and automation (EDA): Electronic component design
and verification
 Geosciences and geo-engineering: Oil and gas exploration and
reservoir modeling
 Mechanical design and drafting: 2D and 3D design and verification,
mechanical modeling
 Defense and energy: Nuclear stewardship, basic and applied research
 Government labs: Basic and applied research
 University/academic: Basic and applied research
 Weather forecasting: Near term and climate/earth modeling
28/11/2012 HPC
6

7. WhoShouldBe UsingHPC?
• High Performance Computing represents a tremendous competitive
edge in the marketplace because it can give users the ability to quickly
model and then manipulate a product or process to see the impact of
various decisions before they are made.
• Existing products and processes can be optimized to reduce cost, and
new avenues of development can be explored at lower cost and with
faster times to market.
• So anyone interested in faster times to solution, better science, informed
decisions, more competitive products can use HPC.
28/11/2012 HPC
7

8. Enter theCommodityCluster
Drawbacks and problems in HPC :
 First the price was steep- in seven figures the entire system was
integrated from top to bottom by a single company leading to high
cost.
As the performance was increasing the cost of designing and
fabricating custom processor and other components was also
increasing.
 considered office processors, no one really considered x86s to be
good at floating point calculations.
 In addition to getting faster, supercomputer-specific processors also
got more expensive to manufacture. Indeed, the expense became so
high that it was no longer feasible to design and fabricate
supercomputer-specific processors.
28/11/2012 HPC
8

9. Hardwareand softwaremovement.
1. As commodity processors from companies like AMD improved,
collections of servers began to outpace even the fastest
supercomputers. The move to commodity hardware had begun.
2. Many of the software tools needed to support a UNIX-like
environment were freely available under the GNU software license.
3. Another enabler of the commodity cluster was the Internet . The
Internet helped in two ways:
• it ignited the demand for Web servers.
• it opened up an avenue for international collaboration, helping solve
many of the early HPC issues.
28/11/2012 HPC
9

10. HOWDoesa Cluster LookLike?
• Head or master node. This node is normally a gateway to a shared
network where users “log in” to the cluster.
• The head node has one or more networks with which it
communicates to the worker (or compute) nodes.
• In terms of networks, a cluster may have as little as one private
network, which is usually Gigabit Ethernet (GigE), although Infini
Band is becoming a popular option.
28/11/2012 HPC
10

11. “Typical”clusterconfigurationModel
Typical Cluster Configuration Contents
• There’s a head node that may
contain a large amount of
storage that is shared via the
network by all the worker
nodes.
• A varying number of worker
nodes communicate with each
other and the head node over
the available private
networks.
28/11/2012 HPC
11

12. REALLIVECLUSTER!!
• A real live cluster often looks
like that shown in figure. The
nodes are placed in standard
equipment racks and usually
require a data center
environment (raised floor,
chilled air, and so on).
28/11/2012 HPC
12

13. TypicalIU ServerV/S Bladed System
Typical “1U server” Bladed system
28/11/2012 HPC
13

14. Categoriesof cluster
• Clusters are often broken into two categories:
1. Capability
2. Capacity
• A capability cluster is designed to handle (or be capable of handling)
large compute jobs that may employ every node in the cluster.
• Capacity clusters, on the other hand, are those that are used to
deliver a certain amount of computing capacity to the end users.
28/11/2012 HPC
14

15. IfYou NeedSpeed!!
• High performance clusters are used where time to solution is
important.
• To increase computing throughput, HPC clusters are used in a variety
of ways.
1 compute farm
2 parallel computing
28/11/2012
HPC
15

16. Computefarm
• The first and easiest way is to allow the cluster to act as a compute
farm.
• Depending on the resources, all the jobs may run at the same time
or some may wait in the queue while other jobs finish.
• This type of computing is local to a cluster node, which means the
node doesn’t communicate with other nodes, but may need high
speed file system access
28/11/2012 HPC
16

17. ComputeFarmProcess
28/11/2012 HPC
17

18. Parallelprogramrunning
• Another way to use clusters is to break large HPC jobs up into smaller
sub-jobs and run each sub-job on different nodes.
• This process is done at the software level
• The most popular method for creating cluster programs is to use the
MPI (Message Passing Interface) library and insert communication
calls in the program..
• The program employs all the processors and must communicate
with the other nodes.
• The parallel program will run much faster than using a single node
28/11/2012
HPC
18

19. ParallelProgram Process
28/11/2012 HPC
19

20. Of Cores, Threads,and Nodes
• Cores are processing units within a modern day processor. A typical
motherboard may hold 1, 2, or 4 processors and is contained in a
node.
• A typical node may have 2 processors, each with 6 cores for a total of
12 cores per node.
• programs on a multi-core node are run in a Symmetric Multi-
Processing (SMP) mode. This designation means that multiple
programs can be running at the same time on various cores.
• Threaded programming is parallel programming designed for a
single SMP node.
28/11/2012 HPC
20

21. AbouttheData!!
• In addition to the parallel computation there may also be a need for
parallel data access.
• If all the nodes were required to read and write data from the same
file at the same time. A bottleneck would develop and limit the speed
of your program.
• For this reason, parallel file systems have been developed. These
solutions often involve additional hardware and software.
28/11/2012 HPC
21

22. ProcessorsAnd Nodes
1)Processor is a workhorse of the cluster.
2)Nodes are each server in a cluster.
3)Cluster nodes may have 8 to 16 cores per nodes.
4)The choice of processor is very important because cluster
installations are often rely on scalable processor performance.
28/11/2012 HPC
22

23. ProcessorsAnd Nodes(contd..)
5) Depending upon the design of the cluster , nodes are:
(i) fat nodes (lots of cores, disk and memory)
(ii) thin nodes ( small cores, disk and memory )
28/11/2012 HPC
23

24. IPMI
• The Intelligent Platform Management Interface specification defines
a set of common interfaces to a computer system. System
administrators can use IPMI to monitor system health and manage
the system.
28/11/2012 HPC
24

25. Co-processors
• They are called array processors or GP-GPU ( General Purpose
Graphical Processing Units ) or video cards . These devices can
accelerate certain types of mathematical operations .
28/11/2012 HPC
25

26. Interconnect
• It is a port that attaches one devices to another. Their performance
is rated by :-
a) Latency [ fastest time in which a single byte can be sent ]
b) Bandwidth [max. data rate ]
c) N/2 Packet Size [throughput curve rise rate]
28/11/2012 HPC
26

27. Interconnect(contd..)
d) Messaging [ tells how much messages per second an interconnect
can send ]
e) Bi-section width [ this rates how well the switch supports many
simultaneous conversations ]
28/11/2012 HPC
27

28. Deploymentof HPC
1) Infini Band (IB)
2) 10Gigabit Ethernet
28/11/2012 HPC
28

29. RememberingTheStorage
• Almost all clusters require HPC storage. The simplest method is to
use the head node as an NFS server and RAID sub-systems.
• HPC applications create large amt of data and it is to be ensured that
an archiving system is available to many data centre.
• Flash modules can be a integrated directly onto motherboards.
28/11/2012 HPC
29

30. RackingAnd Stacking
• Computer nodes can take the form of IU servers or blade systems.
Blade systems often have a slightly higher costs, but offers more
manageability, density and some power and cooling redundancy. It is
to be ensured that all cluster equipment is ‘rack mountable’ and can
be placed in a standard 19 inch equipment racks.
28/11/2012 HPC
30

31. FindingCluster Software
• Linux is by far the most common OS that HPC users choose and other
options include Solaris software .
• The GNU/Linux core is a open software and can be freely copied . It
has allowed HPC users to create apps, build drivers and lots more .
28/11/2012 HPC
31

32. HPC Glue: Cluster Software
• For a successful software running several types of software tasks
include administration, programming, debugging, job scheduling
and provision of nodes.
• Programming is perhaps the most important aspect of cluster and its
tools include MPI (Message Passing Interface ). MPI allows programs
to talk to one another over cluster networks.
• In addition to MPI , programmer requires compilers, debuggers and
propellers .
28/11/2012 HPC
32

33. FileSystems
Now a days parallel file systems developed for HPC systems. They
allow for multi node input and output and are designed to match a
particular cluster. Parallel file system e.g. lustier from Sun
Microsystems.
28/11/2012 HPC
33

34. SharingIs Caring: HPC ResourceSchedules
• The resource scheduler is a critical part of the cluster because it would
be impossible to share resources without same form of load balancing
tools . There are several tools freely available resource schedulers e.g.
Sun Grid Engine from Sun Microsystems , Torque, Lava etc.
28/11/2012 HPC
34

35. Provisioning: CreatingTheCluster
Turning raw hardware into a functioning cluster is called
provisioning and it involves installing and configuring the head
node and worker nodes of the cluster . There are 2 main methods
that are used :
I. Local
II. Remote
28/11/2012 HPC
35

36. ClusterToolKits
• The free available cluster distributions that provide turn key
provisioning ready to run software are :-
A. Sun HPC Software , Linux Edition
B. Rocks Clusters
C. Oscar
28/11/2012 HPC
36

37. YOU
THANK
28/11/2012 HPC
37