MPI _ HPCC Based on zju-309 virtualization LaB _ Image

Source: Internet
Author: User

MPI Based on zju-309 virtualization LaB _ ImageFirst, copy the number of required images through the original image (. IMG. Swap. cfg) as needed. Perform the following steps to configure the image.

1. Physical file configuration (*. cfg)

  • Memory = 256
  • CPU apsaradb for memory = 256
  • Image address

Disk = [

8 'file:/share/nfs04/xen/raw-img/node11.swap, xvda1, W ',

9

10 'file:/share/nfs04/xen/raw-img/node11.img, xvda2, W ',

11]

  • Name = 'zlwnode11'
  • Mac cannot have the same name. You can modify the name as needed. Note that the first two cannot be modified.

VIF = ['mac = 00: 16: 3E: 4f: F8: 61, bridge = br0']

Note: br0 is the name of the Physical Network Bridge.

 

2. Inside the VM

  • Modify the IP addresses of all nodes;
  • Modify the MPD host configuration file

Root @ zlwnode1 :~ # Vim/root/MPD. Hosts

For example, vc1_1 is the host name.

1 vc1_1

2 vc1_2

  • Modify hosts

Root @ zlwnode1 :~ # Vim/etc/hosts

Example:

1 10.214.15.71 vc2_1

2 10.214.15.72 vc2_2

3 10.214.15.73 vc2_3

4 10.214.15.74 vc2_4

5 10.214.15.75 vc2_5

6 10.214.15.76 vc2_6

7 10.214.15.77 vc2_7

8 10.214.15.78 vc2_8

9

10 # the following lines are desirable for IPv6 capable hosts

11: 1 localhost ip6-localhost ip6-loopback

Ip6-localnet 12 fe00: 0

Ip6-mcastprefix 13 ff00: 0

14 ff02: 1 ip6-allnodes

15 ff02: 2 ip6-allrouters

16 ff02: 3 ip6-allhosts

 

  • Copy the MPD. Hosts and hosts files to the symmetric node.

 

3.configuration file configuration (hpccinf.txt)

 

8 device out (6 = stdout, 7 = stderr, file)

6 # of problems sizes (N)

4000 4000 NS(Problem scale)

 

 

1 pmap process mapping (0 = row-, 1 = column-Major)

8 # of process grids (P x q)(Number of threads)

8 1 4 Ps

1 8 2 Qs

 

 

P × q = number of system CPUs = number of processes.

P ≤ q; in general, the value of P should be smaller as possible, because the column-oriented traffic (the number of communications and the amount of data) is much larger than that of horizontal communication.

P = 2n, that is, it is better to select the power of 2 for P. In hpl, the binary Exchange (Binary exchange) method is used for column-oriented communication. When the number of column-oriented processors P is a power of 2, the performance is optimal. For example, when the number of system processes is 4, it is better to set p × Q to 1 × 4 than 2 × 2.

 

4. Check whether the configuration is successful.

Check the error through the MPD. Hosts file
Mpdcheck-f mpd. Hosts
If no error occurs
Mpdcheck-f mpd. hosts-ssh

5. Run the cluster system

Mpdboot-N 2-f mpd. Hosts
The-N 2 parameter specifies the number of machines to be started, and-f mpd. hosts specifies the number of machines to run through MPD. hosts.
Mpdtrace

Check whether all threads are started successfully
Mpdallexit

Exit all threads

 

Run the command: (Note that the command execution location is hpccin.txt and hppccout.txt)

Start thread

Root @ zlwnode1 :~ # Mpdboot-N 8-f mpd. Hosts

Root @ zlwnode1 :~ # Mpdtrace

Run:

Root @ zlwnode1 :~ /Hpcc-1.3.1 # mpirun-NP 8/root/hpcc-1.3.1/HPCC

 

Running on the physical machine: Watch-D-N 2 XM list to monitor whether the virtual machine is running normally ---

 

 

Attachment:

  • Input File

Hpccinf input file

Maid input file

Innovative computing laboratory, University of Tennessee

Hpl. Out output file name (if any)

8 device out (6 = stdout, 7 = stderr, file)

6 # of problems sizes (N)

4000 4000 NS

2 # of NBS

80 100 NBS

1 pmap process mapping (0 = row-, 1 = column-Major)

8 # of process grids (P x q)

8 1 4 Ps

1 8 2 Qs

16.0 threshold

3 # of panel fact

0 1 2 pfacts (0 = left, 1 = Crout, 2 = right)

2 # of recursive stopping criterium

2 4 nbmins (> = 1)

1 # of panels in Recursion

2 ndivs

3 # of recursive panel fact.

0 1 2 rfacts (0 = left, 1 = Crout, 2 = right)

1 # of Broadcast

1 bcasts (0 = 1rg, 1 = 1rm, 2 = 2rg, 3 = 2rm, 4 = LNG, 5 = LNM)

1 # Of lookahead depth

1 depths (> = 0)

2 swap (0 = bin-exch, 1 = long, 2 = mix)

64 swapping threshold

0 L1 in (0 = transposed, 1 = No-transposed) Form

0 U in (0 = transposed, 1 = No-transposed) Form

1 equilibration (0 = No, 1 = yes)

8 memory alignment in double (> 0)

##### This line (No. 32) is ignored (it serves as a separator ).######

0 Number of additional problem sizes for ptrans

1200 10000 30000 values of N

0 Number of additional blocking sizes for ptrans

40 9 8 13 13 20 16 32 64 values of Nb

 

 

  • Script for replacing the configuration file

 

1 conflist = 'seq 1 16'

2 For conffile in $ conflist

3 do

4 confile = node1 _ $ conffile

5 nodename = vc1 _ $ conffile

6 IP = $ (conffile + 82 ))

7 Echo $ IP

8 IPaddress = 10.214.15. $ IP

9 echo $ confile

10 memory_cof = 512

11 echo $ confile

12 sed-I "s/memory =. */memory = $ memory_cof/".../$ confile. cfg

13 sed-I "s/node8.swap/$ confile. Swap/" ../$ confile. cfg

14 sed-I "s/node8.img/$ confile. img/".../$ confile. cfg

15 sed-I "s/name =. */name = '$ nodename'/".../$ confile. cfg

16 sed-I "s/MAC =. */MAC = 00: 16: 3E: 4f: F1: $ IP, bridge = br0']/".. /$ confile. cfg

17 Mount-o loop ../$ confile. img/mnt

18 sed-I "s/address. */address $ IPaddress/"/mnt/etc/Network/interfaces

19 sed-I "s/zlw. */$ nodename/"/mnt/etc/hostname

20 cp mpd. hosts/mnt/root/MPD. Hosts

21 CP hosts/mnt/etc/hosts

22 umount/mnt

23 done

 

 

Problem feedback: azhuliangwei@qq.com

Document update address: http://note.sdo.com/u/1188891469#/n/6TgJd ~ Jsinfowe0k4001gd

 

 

 

 

 

 

 

 

 

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.