Can you help me solve problem 8 and 9 I am learning it as I go for Parallel comp
ID: 3857318 • Letter: C
Question
Can you help me solve problem 8 and 9 I am learning it as I go for Parallel computing It's for computer science Can you help me solve problem 8 and 9 I am learning it as I go for Parallel computing It's for computer science 1-8. A company is proposing an IPv8 format using 512 bits. Do you think this is justified? Explain. 1-9. It is possible to construct a system physically that is a hybrid of a message-passing multicomputer and a shared memory multiprocessor. Write a report on how this might be achieved and its relative advantages over a pure message-passing system and a pure shared memory system. 10. (Research project) Write a report on the prospects for a trulv incrementallv scalable cluster comnuterExplanation / Answer
1-8.Answer.
-----------------------------------------------------------------------------------------
IPv8 is called the "Global Telecommunication Community".
By using the IPv8 we can eliminate the software which are used for telephone replacement.
We can also obligate to have number beyond the 15 digits
As far as I know the IPV8 uses the same chip as the IPV6 at it's latest revision and has
roughly the same form factor, so it is just P4U attempt of making the IPV6 go away fast.
So if the company is proposing the IPv8 format its fine to go with it. If the purpose is for
Global Telecommunication.
1-9.Answer
---------------------------------------------------------------------------------------
If the system is physically that is a hydrid of a message passing multicomputer and a shred memory multi-
processor then Message Passing Interface can lead it to the point.
Where MPI is a library for parallelism distributing memory having the cluster.
In a simple way the bunch of PC's together with ethernet cable. MPI allows programs
on one node to send data to another or conversly receive.
Why because each cluster node has its own memory and can not directly access the memory
of another node. MPI is very general in that point of matter.
A pure shared memory system
The cost of synchronising access to shared memory is very high when done simply and very hard to do correctly
and efficiently.
You cannot simply write to shared state without some concurrency control as race conditions and other
artefacts are certain to occur.
Therefore, having many processors waiting to modify shared state means that most of them must be
waiting while only one is actively modifying state. This is not especially efficient.