Computers in the acoustics workgroup

1 LINUX SYSTEMS AT NTV

The computer systems are given in the table below. Cowry and Nerite are the main computer server.
Cowry is the main ssh gateway for the systems (and thus much heavier loaded than nerite, use our script ztop to see load on each system).
Nerite is license server for intel fortran compiler.
Cameo does admin stuff and mounts all our Jetstor systems. Jetstor one (/net/cameo/js) has 8TB of storage and Jetstor two (/net/cameo/js2) has 12 TB of storage.
Both Jetstor systems and /mpl are using raid.

Computer	CPU	RAM/SWAP	Location
Cowry	Xeon 4 duel processors= 8 CPU/ 3Ghz	32GB/30GB	Campus
Nerite	Xeon 4 duel processors= 8 CPU/ 3Ghz	32GB/30GB	Campus

2 GUIDE LINES FOR CPU USAGE

The optimal load on Nerite and Cowry are 6 one-cpu jobs, whith more load it tend to just become more swapping.
For cowry and nerite we have the following guidelines:

1) No more than two long jobs per user in total during work hours.
2) In off hours, it is more flexible, but likely no more than 4 one-cpu jobs in total.
3) With light loading you can break the rules, but be ready to kill jobs if needed.

Some advice:
a) The load on both systems can be monitored by using ztop or the simpler nodeload.
b) If you have many jobs to run, put them into a file and executed sequentially.
c) A job should not take more than a day to run. Or it should be split into sequences
that can be tested. That way results can be tested before continuing.
d) For long jobs write the results out with regular intervals (e.g. once an hour).

Cameo does admin stuff and mounts all our Jetstor systems. Jetstor one (/net/cameo/js) has 8TB of storage and Jetstor two (/net/cameo/js2) has 12 TB of storage.
Both Jetstor systems and /mpl are using raid.
The maximum RAM memory for a job is 10 GB of Cowry and 20 GB on Nerite. If you run that large jobs, be careful how many you have running as we only have 32 GB RAM on each system. all jobs are niced max after running 20 min CPU.

Computer	max Ram	Max Matlab Thread	Max CPU time
Cowry	10 GB	2	Infinite
Nerite	20 GB	8	Infinite

3 DISK SPACE AND DATA DIRECTORIES

You can see how much disk space is used by the command
    df -h
Look for /mpl, /cowry0, /nerite0, /js, and /js2. Awoid filling up the disk completely.

You can see how much disk space you use in a directoy and subdirectories
    du -sh *

To write a large file it takes a factor 10 longer on a remotely mounted disk than on a local one.
Thus, for a disk-intensive job, it is recommended running it on a local disk, ie. /cowry0 or /nerite0.

Home directories:
Users are not supposed to use their home dir for data storage as it can only hold max 5GB/user,
This is controlled by setting up quotas for each user. You can see your disk quota. Use
    quota -sQ
To see your limits.

Jetstor:
Our two Jetstor systems contains most of our data storrage.
Experimental data are stored in /net/cameo/js2/data and containg 12 TB of data

Users directories are on /net/cameo/js/data and contain 8 TB of data

A subdir /net/cameo/js/data/$USER will be created if needed with an appropriate quota (ask mpladmin).
It is usually convenient to make a softlink to that directory in your home dir:
    ln -s /net/cameo/js/data/gerstoft js

Quotes are installed on /js with 700GB soft limit and 1TB hard limit.
The hard limit defines the absolute maximum disk space available to a
user. The soft limit is more lenient, in that the user is only warned
when he exceeds the limit. He can still create new files and use more
space for a grace period of 7 days.

On 9 Nov 08: cameo:/js/data        7.6T 4.7T 2.8T 63% /.automount/cameo/root/js/data
On 5 Dec 08: cameo:/js/data      7.6T 5.4T 2.1T 73% /.automount/cameo/root/js/data
ON 9 July 09: cameo:/js/data        7.6T 5.9T 1.6T 79% /.automount/cameo/root/js/data
ON 23 Oct 09 cameo:/js/data        7.6T 6.9T 592G 93% /.automount/cameo/root/js/data

Nerite and Cowry
Each has one 0.5 TB disk ( /net/cowry/cowry0 and /net/nerite/nerite0)

Backup/Mirroring
The home directory (/mpl ) and /nerite0, /cowry0 are mirrored each night.
The jetstor is not mirroed but because of their size. However, if a single disk fails it should recover automatically due to the raid6

4 GRAPHICS

While most plots is produced in matlab, users can display plots from a remote system using:

eog     works with {ani,bmp,gif,ico,jpeg,pcx,png,pnm,ras,svg,tga,tiff,wbmp,xpm}
ggv                         {ps,PS,eps}
xpdf                        {pdf}
acroread                  {pdf}

For generating graphics it is usually easier to do it on your laptop. We have installed xfig and imageshark. Imageshark is useful for making movies, see
http://www.imagemagick.org/

5 LONG MATLAB JOBS

Before you start a long job remember to make many small tests--- If is a lot easier to find bugs in small runs!

For a long job you can save the output to mat file and then later read that into matlab for post-processing (recommended). Alternatively you you can print your figures to a file.

If a user needs to run long jobs on cowry or nerite he can start them in background and then log out from the console.
This script (matlabrun) shows how to do it for matlab:

    matlab -nodisplay << INN >> ./list.out
    MyMatlabProgram
    quit
    INN

and then (with script and matlab program in same directory)
    matlabrun &

Diana suggests:
    nohup nice +19 matlab -nodisplay -nodesktop -nojvm -nosplash < MyMatlabProgram.m >& ./list.out &

6 USER ACCOUNT SETUP

User environment is set up by sourcing system-wide /usr/local/etc/cshrc and /usr/local/etc/login by $USER/.cshrc and $USER/.login.

Users can overwrite system-wide settings or add their own in
~/.cshrc.custom.linux
~/.login.cstom.linux
~/.aliases
~/.aliases.linux

Users are supposed to put their own binaries in subdir 'bin' in their home dir, this dir precedes system dirs in case of name conflict.

By default user's $path is set up to:

/usr/local/intel/fce/10.1.015/bin:.:/mpl/$USER/bin:/mpl/$USER/x86_64/bin:/usr/local/bin:/usr/X11/bin:/usr/bin:/b
in:/usr/local/matlab/bin:/usr/local/gmt/bin

7 Username and password

At UCSD your username is usually initial+lastname (e.g pgerstoft). But sometimes a truncated 8 character version (e.g. pgerstof) of this is used. Sometimes you can also choose an alias (e.g. gerstoft).

Password are stored centally across some systems as shown here:
http://blink.ucsd.edu/Blink/External/Topics/How_To/0,1260,24771,00.html

The linux system has their own username/password.

8 NTV PRINTERS

1. black and white HP LaserJet 4250
printer name on linux systems:
single-sided print - lwwaklabs                           double-sided print - lwwaklab
DNS Name:    mpl-lwwaklab.ucsd.edu             IP Address: 172.16.128.16

2. color HP Color LaserJet 4700
printer name on linux systems:
single-sided print - lwwaklabds                       double-sided print - lwwaklabd
DNS Name:    mpl-lwwaklabd.ucsd.edu           IP Address: 172.16.128.18

9 Internet

ssh
netbrowser: Firefox
ftp client is also available. If transferring binaryfiles you have to set the bin flag. Ulike on a Mac.
curl is a command line tool for transferring files with URL syntax, supporting FTP, FTPS, HTTP, HTTPS, SCP etc. It works better than classical ftp.

10 NOTE FROM MPL SYSADMINS

All linux systems in ntv including desktops are shared computing resources which can be used by any user who has account on those system, and since some users might login remotely, please do not reboot your linux desktops in attempt to fix problems w/o contacting mplhelp. We will then try to resolve this problem and notify other users about reboot if needed.

Please log out from your desktops when you are not using them so we can do system maintenance from the console and upgrade/reboot systems when needed.

You can send mail to mplhelp@mpl.ucsd.edu or call:
Diana: x45804 Irina: x45503 Rob: x45954

11 SAGA benchmark:

cd saga/examples/ramgeo/vertical/
saga tc1v1lay ramgeo

saga/ramgeo running tc1v1lay from the inversion workshop
(10000 forward models) on one CPU.
CPU Times:

astartes                  9603.1   25 jan 01
turrid                     23821.9
heart                     18724 s
heart (f77 -fast )       14812 s
heart (f77 -fast Solaris 2.8) 5297.4 (27 Dec 02)
ark (David B) f77-fast 26094 s
sundial                  13046.9   25 jan
baby                    2368.8 s (g77)
baby                    1771 s ifc (22 Jan 03)
baby                    1434.4 (ifc -O -axW -zero -w, 2 Oct 03)
baby                    1475 (ifort -O -axW -u   -w -posixlib -Vaxlib -CB, Feb 06)
pupa                    1290.6 (ifc -O -axW -zero -w, 2 Oct 03)
occam (Mac G4 1.5GHz laptop) 4159 s (xlf -O )
G5 2 GHz Mac            1700    (xlf -O)
chenfen laptop(Mac intel ) 1237 (ifort aug 06)
Peter's laptop (Mac pro book 2.33 Ghz Core 2 duo ) 675 (ifort 9)
Peter's laptop (Mac pro book 2.33 Ghz Core 2 duo ) 637 (ifort 10.1) 27Mar 07
Mac workstaion 8GB, 3GHz 2xdual Xeon       531
Nerite   2Xquad 568 (ifort v10) 2008
Nerite   2Xquad 575 (ifort v11) June2009
Caglars laptop (Mac pro book 2.5 Ghz Core 2 duo , ifort 11) 521 June09
Peter laptop (Mac pro book 3.06 Ghz Core 2 duo , ifort 10) 434 June09
Peter laptop (Mac pro book 3.06 Ghz Core 2 duo , ifort 11) 425 July09
Peter laptop (Mac pro book 3.06 Ghz Core 2 duo , ifort 11.1 Mac10.6) 418 Oct09

12 WWW upload

I develop my webpages locally and then upload it all, eg:

scp -r *html pgerstoft@mplwd.ucsd.edu:WWW

where the WWW is a
The password on mplwd is the netop passwd.