BUILDING of 3 R'sCLUSTER
PARALLEL COMPUTER
Presented by:
Yash Gurung, ICFAI UNIVERSITY.Sikkim
Introduction:
 Parallel computing may be defined as the simultaneous use of
several processors to execute a computer program.
 There two types of parallel architectures
 Shared memory multiprocessor(SMP)
 Cluster Parallel Computer
Figure 1: Cluster Parallel computer
Cluster:
 A cluster is simply a “stack” of processor machines usually
mounted in a rack.
 The stack is configured such that one machine is designated
the head node and a file server node to all the other
machines.
 Client nodes technically known as slave nodes are connected,
via an internal network
Figure 2: Cluster Computer Architecture
Literature Review
 Bewulf Cluster Project: was started at the Center of
Excellence in Space Data and Information Sciences NASA in
early 1994.
Hardware Configuration specification:
 One Master Node:
 Two Pentium III 1GHz CPU, 512M RAM, three 3Com 3c905c
Ethernet cards, one 30G bytes hard disk, one floppy, one VGA card,
one monitor
 One NFS and NFS Root Server Node:
 Two Pentium III 1GHz CPU, 512M RAM, two 3Com 3c905c
Ethernet cards, one 30G bytes hard disk, one floppy, one VGA card,
one monitor
 11 Client Nodes:
 One Pentium III 1GHz CPU, 512M RAM, two 3Com 3c905c
Ethernet cards, no hard disk, one floppy, one VGA cards (for
debugging)
 2 Hubs
 D-Link DES-1016R, D-Link DFE-916DX
Software Configuration specification:
 Operating system: Red Hat 6.2
 Network booting: Etherboot-4.0
 Parallel computing: Message Passing Interface mpich-1.2.1
Figure 2: Example of Cluster Computer SUGAR cluster.
SUGAR CLUSTER:
 The Syracuse group has constructed a 320 CPU node parallel
computer called the SUGAR cluster. It consists of
 Head nodes: 6 Sun X2200-M2 computers, each containing
two Dual-Core AMD Opteron 2.6 GHz processors
 Compute nodes: Eighty servers each with one quad core
Intel 2.4 GHz Xeon CPU and 8 Gb of RAM provide a total
of 320 CPU cores and 640 Gb of RAM for the cluster
 Cluster storage is provided by four file servers, each with 24
Tb of disk storage
 The compute nodes and head nodes are connected via copper
gigabit Ethernet (1000baseT)to a switch.
ALICE CLUSTER:
 a cluster of 64 dual-processor Pentium Pros running at 200
MHz
 All PCs are connected through a central Fast Ethernet switch
 In addition, there is a master node, a file server, and 4
development nodes.
Figure 3: ALICE CLUSTER
Justification:
 The three R’s:
Reuse.
 Reduce.
 Recycle.

 Building of a computer by reusing the hardware components
of the unused computers
Figure 4: Possible Configuration
Figure 5 : Possible Configuration
Methodology:
1. Survey and Collection of functional hardware and both
software components from various sources.
2. Assembling of the components into a functional unit.
3. Designing a layout for the final product.
4. Loading of necessary Softwares.
5. Analysis and testing of the product.
Expected results:
 High Performance.
 Expandability.
 High Throughput.
 High Availability
Conclusion:
 It is certain that building a new computer by recyling the out
of use hardware product, will invite lot of problems
regarding the various issues that are related with the
designing, implenting etc.
 But it is still encouraging to construct a working computer
then not to use it all.