NEW TESTS BASED ON PARALLEL COMPUTING DURING MIGRATION

I took the same tests on the VMs where they are in the process of migration, they worked however the processing time of the running program is not considerable longer, and some of them are exactly the same. I am confused aboutnthe performace difference, and it should be some thing there to show the delay due to copying memory files and page files.
I thought about another situation which would be interesting and revelant to this project---zip and unzip files. I just took three times to show the performance difference between running isolated and running during migration. These are some data:
zip file No.1 size is around 2500MB, it has single file inside.
zip file No.2 size is around 4GB, it has over 200 files inside.
folder No.1 has over 100 files, the total size is around 2GB.

OS VMID TIME(ISO)   TIME(MIG)   FILE PROCESS
Ubuntu 1 65s       75s        zip1 UNZIP
ubuntu 1 97s        132s      zip2 UNZIP
ubuntu 1 116s       180s      folder1 ZIP    

DATA FROM NEW TESTS ON VMS

I designed a couple of programsto fit the aigs of parallel computing.
There are three small programs/functions, which are:
1. Array addition.(AA)
2. Matrix multiplication (Ax)
3. Matrix multiplication (AB)

The first step I took is to run the programs on those VMs before any migrations, and collect the data as following:
VMNAME VMID VCPU RAM/GB PROGRAM  SPENT TIME/10^-3s(AVERAGE)
ubuntu 1 2 1.5 AA 781
ubuntu 1 2 1 AA 811
ubuntu 1 2 0.5 AA 2132
ubuntu 1 1 1.5 AA 955
ubuntu 1 1 1 AA 1205
ubuntu 1 1 0.5 AA 3577
deepin 2 2 1 AA 890
deepin 2 1 1 AA 1239
ubuntu 1 2 1.5 AA  713
ubuntu 1 2 1 AA  1030
ubuntu 1 2 0.5 AA  917
ubuntu 1 1 1.5 AA  1208*
ubuntu 1 1 1 AA  1046
ubuntu 1 1 0.5 AA  3526
deepin 2 2 1 AA  882
deepin 2 1 1 AA  1200


VMNAME VMID VCPU RAM/GB PROGRAM  SPENT TIME/10^-3s(AVERAGE)
ubuntu 1 2 1.5 Ax 3928
ubuntu 1 2 1 Ax 4124
ubuntu 1 2 0.5 Ax 11730**
ubuntu 1 1 1.5 Ax 9875
ubuntu 1 1 1 Ax 9583***
ubuntu 1 1 0.5 Ax 23737
deepin 2 2 1 Ax 4359
deepin 2 1 1 Ax 6338
ubuntu 1 2 1.5 Ax 2399
ubuntu 1 2 1 Ax  4850
ubuntu 1 2 0.5 Ax  5674
ubuntu 1 1 1.5 Ax  4755
ubuntu 1 1 1 Ax  7136
ubuntu 1 1 0.5 Ax  9090****
deepin 2 2 1 Ax  3644
deepin 2 1 1 Ax  7653

VMNAME VMID VCPU RAM/GB PROGRAM  SPENT TIME/10^-3s(AVERAGE)
ubuntu 1 2 1.5 AB 6584
ubuntu 1 2 1 AB 7099
ubuntu 1 2 0.5 AB 7346
ubuntu 1 1 1.5 AB 6422
ubuntu 1 1 1 AB 6510
ubuntu 1 1 0.5 AB 9917
deepin 2 2 1 AB 6139
deepin 2 1 1 AB 8677
ubuntu 1 2 1.5 AB  5931
ubuntu 1 2 1 AB  6649
ubuntu 1 2 0.5 AB  7316
ubuntu 1 1 1.5 AB  8227
ubuntu 1 1 1 AB  9100
ubuntu 1 1 0.5 AB  9730
deepin 2 2 1 AB  5657
deepin 2 1 1 AB  7892

NEW TESTS BASED ON PARALLEL COMPUTING

After talked with Dr. Parimala, we figured out that,for the tests over the migration procedure, it is better to have parallel data structure and computation. There are a few options for me:
1. Array addition.
2. Matrix multiplication (Ax)
3. Matrix multiplication (AB)
With these specified data structure, the problem of data requirement will not be any problem. And at the same time, in order to enlarge the size of the program, I may use double or long variables.
    I spent a couple of days to design and complete the test program with some basic tests. Now it is ready for running it on VMs