Centos communication network. The latency is measured

Centos 7 installation completed, Docker old
version uninstalled, and new version is installed successfully by the record of
logbook and with the help of these screen shots shown above.

Compilation of HPC is for Vcc. Extracting that
to the cluster.it brings together 7 popular benchmarks that access CPU, memory
and networking performance. Reference: unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

Best services for writing your paper according to Trustpilot

Premium Partner
From $18.00 per page
4,8 / 5
4,80
Writers Experience
4,80
Delivery
4,90
Support
4,70
Price
Recommended Service
From $13.90 per page
4,6 / 5
4,70
Writers Experience
4,70
Delivery
4,60
Support
4,60
Price
From $20.00 per page
4,5 / 5
4,80
Writers Experience
4,50
Delivery
4,40
Support
4,10
Price
* All Partners were chosen among 50+ writing services by our Customer Satisfaction Team

1.   HPL

2.   DGEMM

3.   STREAM

4.   PTRANS

5.   RANDOM
ACCESS

6.   FFT

7.    B-eff     

 

Hpl

It is
used to measures the floating-point rate of

Execution
for the solution of a linear system. It is the most popular benchmark of CPU
execution

Performance
and is used to rank the TOP500.

It is
also the most complex test, both in terms of computational

Difficulty
and installation/configuration, but we will come back

to
this later on. Reference: unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

 

DGEMM

DGEMM measures the floating-point execution
rate for double

Precision matrix multiplication.

This problem is similar to the MPI matrix
multiplication

Example in the lab. It is easier to understand
how it works than

the LINPACK, but is less computationally
expensive. Reference: http:unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

 

STREAM

Stream benchmark is a simple benchmark program,
which is used to measures sustainable memory bandwidth in mb/s and the
corresponding computation rate for simple vector. Reference: http://icl.cs.utk.edu/hpcc/links/index.html

 

PTRANS

It measures the rate of transfer of large data
arrays between multiple processors memory. https://unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

Random Access

It measures the rate at which random segments
of memory can be updated.

A worst-case scenario is represented by the
random access where the data must be written to memory is non-contiguous;
however this is the exact application that random access memory is designed and
optimised for. https://unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

 

FFT

Floating point is measured by the FFT.the rate
of execution of a double precision one-dimensional discrete Fourier transform.
FFT is not expensive as the LINPACK is. https://unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

B_eff

The Effective Bandwidth benchmark is an
assessment of

Latency and Bandwidth of the communication
network.

The latency is measured using a small message
size and the

bandwidth is measured using the largest
possible message size.

Both tests are performed during
non-simultaneous and

Simultaneous communication, using basic MPI
routines

(send/recv and scatter/gather).reference: https://unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

 

 

 

 

CLOUD COMPUTING

Cloud services types mainly consists of Saas,Paas,Iaas

SaaS
(software as a service):

Benefits of SaaS is to sign up and start rapid
using the apps.

Data and Apps can be easily accessed from any
computer.

If the computer is broken, down. Data will be
safe still because data is in the cloud not in the computer.

Reference:

https://www.ibm.com/cloud/learn/what-is-cloud-computing

 

PaaS

It gives a environment of cloud with each and
everything which is essential for the supporting of cycle of manufacturing and
delivering web-based (cloud) apps except the price and thinks to buy and its
managing hardware and software.

Advantages
of PaaS

It develops the applications and get to the
market as early as possible.

Assigns new internet apps to a cloud within
minutes.it decreases the complexion with middleware as a service. reference: https://www.ibm.com/cloud/learn/what-is-cloud-computing

 

IaaS

Infrastructure as a service gives computing
services to a organizations such as memory storage, networking facility. Advantages
of IaaS are

There is no need to spend money in own
hardware.

It is private or public.

Its provisioning is faster.reference: https://www.ibm.com/cloud/learn/what-is-cloud-computing

 

Open stack

Set of software tools for making and managing cloud-computing
systems for public or private computing clouds is an open stack. Open stack allows
the user to send virtual machines and different examples that handles different
assignment to deal with the cloud. The main thing is that open stack is a free
source programming which means that anybody or individual who selects to get
the source code, needs any improvement or changes which the user wants or
requires and freely share these progressions pull out back to the
group.reference: https://opensource.com/resources/what-is-openstack

There are different types of open stack cloud
users like some of them are administrator, user and operator. Everyone have a
security access levels difference.

Open stack components:

reference: https://unilearn.hud.ac.uk/webapps/blackboard/execute/displayLearningUnit?course_id=_29718_1&content_id=_2309377_1

 

Nova is an essential part of the open stack
that gives the help to the administration of virtual machines occasion.

 Swift is
used for the files storage in the cloud and images and the backup files.

Cinder is a block storage element. It manages
the life cycle of the block devices.

Neutron is networking component.it is
responsible to provide the different services to the users of cloud for example
IP address, DNS, DHCP.

HORIZON is dash board behind the open
stack.open stack has only this graphical interface.

KEYSTONE is responsible for the identity
services for the open stack.

GLANCE is a image service provider to the open
stack.

SAHARA is used to process the data
services.reference: https://opensource.com/resources/what-is-openstack

https://unilearn.hud.ac.uk/webapps/blackboard/execute/displayLearningUnit?course_id=_29718_1&content_id=_2309377_1

 

Figure 13:open stack dashboard

 

Figure 14:instances

 

Figure 15:scheduling instances

Figure 16:centos  terminal

 

 

Figure 17:command running

 

 

 

Figure 18:entering the password for cloud

 

Figure 19:uploading html webpage to the cloud

Different
platforms are different in their performances, the possible bottleneck exist in
the various benchmarked systems, the most important part of the procedures
particularly depends on that information obtained from the test cases. These
types of analyses control those area of the bottlenecks which hits those
execution glitches signifies destruction of the developing in the amount from
claiming utilization.

FUTURE WORK

Designing
of clusters are used to keep the availability of applications, instead of keeping
the data available. Protection against viruses, files corruption, and more
problems facing with the files. Companies needs more strong protection of their
data and the recovery for the data lost. Services of cluster and the service
administration behind the clusters relies on the perfect applications and the
correct services which works properly.

However,
it will be important to deploy a future work to run the data which might give
results to a client of cluster. And the particular improvement which will be
utilized for the improvement of performance.

Reference v.sriram Kumar, R.a.m.k.u.m.a.r. (2014).
cluster computing. international
journal of computer science and information technology, 2 , 319-324.
Retrieved from www.researchpublish.com.

RESULTS AND DISCUSSION

Some
of the results from the hpc are shown below in the figures that shows the
progression of the work during the lab sessions. When we allocate the
software’s into the system and get the understanding of the Linux, centos 7
installation how to configure the network. We work extra hours in the uni lab
to understand the clusters and it’s a success to make a cluster within the
university and making of web page and uploading on the cloud.

Detailed results are attached with the report
to understand properly.

CONCLUSION

Computer
cluster is well famous now a day in everyday of modern age of technological
improvement, for the researchers and the most imp for the students who are
learning the way of improvements in computer industries. Cluster is very useful
for facing the challenges and the number of problems come into the computing aspects.
The main conclusion for this report is that it is easy to achieve the building
of benchmarking clusters. Cluster benchmark is outlined and information
considerations are achievable for the results.

 

REFERENCES

v.sriram kumar, R.a.m.k.u.m.a.r. (2014).
cluster computing. international
journal of computer science and information technology, 2 , 319-324.
Retrieved from www.researchpublish.com.

1.   https://unilearn.hud.ac.uk/webapps/blackboard/execute/displayLearningUnit?course_id=_29718_1&content_id=_2309377_1

2.   https://opensource.com/resources/what-is-openstack

3.   https://unilearn.hud.ac.uk/webapps/blackboard/execute/displayLearningUnit?course_id=_29718_1&content_id=_2309377_1

4.   https://unilearn.hud.ac.uk/webapps/blackboard/execute/displayLearningUnit?course_id=_29718_1&content_id=_2309377_1

5.   https://www.ibm.com/cloud/learn/what-is-cloud-computing

6.   https://unilearn.hud.ac.uk/bbcswebdav/pid-2177647-dt-content-rid-3396507_1/courses/NME3502-1718/hpcc%281%29.pdf

7.   http://icl.cs.utk.edu/hpcc/links/index.html

8.   : http://icl.cs.utk.edu/hpcc/

9.   http://www.cloudbus.org/papers/ic_cluster.pdf