History diff of GetStartedwithFLEX(Open) (No. 2)

• History
• View the diff current.
• View the source.
• View the history.
• Go to GetStartedwithFLEX(Open).
  • 1 (2014-03-20 (Thu) 12:21:00)
  • 2 (2014-03-20 (Thu) 12:21:49)

• The added line is THIS COLOR.
• The deleted line is THIS COLOR.

[started by Hasdeo, 2013.2.25] [edited by Nugraha, 2013.10.22] 
[started by Hasdeo, 2013.2.25] [edited by Nugraha, 2014.3.20] 

#contents

** What is FLEX? [#flex]
FLEX (http://flex.phys.tohoku.ac.jp) is the Saito-lab server computer.  It contains all the files of Saito-lab's staffs and students, including:
+ research-related files: fortran program, scientific paper, etc.
-- please make the following directories in your HOME:
 $ mkdir for tex espresso vasp
 $ cd tex
 $ mkdir paper thesis
-- (for: fortran program, tex: latex files, espresso: quantum espresso input files, vasp: VASP input files)
+ webpages:
 $ cd ~/public_html
-- you can put your website there.
+ e-mail: http://flex.phys.tohoku.ac.jp/webmail, you can login with the same password as FLEX account.

Important note: DON'T forget to change your password if it's just created for the first time:
 $ yppasswd
-- just type the yppasswd command and follow the instructions.
-- do not use a simple "passwd" command because we are using NIS for communicating with cluster machines.

Since FLEX should run heavy load information, very big Fortran program (~ more than 5 minutes running) SHOULD NOT be executed under FLEX. We have to run the program in the other computers whose specialization is for running a heavy program. We call these computers as a TUBE system.

** TUBE system [#tube]
TUBE is divided into 4 sub-systems from different generation (from the oldest to the latest), i.e.
+ TUBE20 series 
+ TUBE30 series (tube30, tube31, tube32)
+ TUBE40 series (tube40,tube41, tube42)
+ TUBE50 series (tube50, tube51, each has 8 cores)
+ TUBE60 series (so far only tube60, it has 40 cores)
Note: TUBE20s and TUBE30s are now retired. TUBE40s are shutdown.

** How to connect to TUBE [#tube-connect]
For example if you want to connect to tube50, 
+check whether a TUBE machine is on or off, by using a command
 $ /rsaito/bin/ww
or if you are lucky enough, you can simply type
 $ ww
+from FLEX terminal, then  use a command 
 $ ssh tube50
or 
 $ t50

** How to submit a job in TUBE [#submitjob]
+Check whether anybody else is running the program or not, use a command
 top
+Then you will see
 top - 11:55:16 up 75 days, 59 min,  3 users,  load average: 1.00, 0.99, 0.89
 Tasks: 157 total,   2 running, 155 sleeping,   0 stopped,   0 zombie
 Cpu(s): 12.5%us,  0.0%sy,  0.0%ni, 87.5%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Mem:  16408160k total,  6641392k used,  9766768k free,   168860k buffers
 Swap: 16746492k total,   539436k used, 16207056k free,  5936008k cached
 PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND            
 31824 nugraha   20   0 75808  59m 1668 R  100  0.4  27:45.75 cpbandgap.out      
    1 root      20   0 24780 1840  708 S    0  0.0   0:02.10 init               
    2 root      20   0     0    0    0 S    0  0.0   0:00.26 kthreadd           
This means Nugraha-san is now using tube50. Please check the ''load average''. The load average should not be more than ''7''. If that is the case, you should go to the other computers.
+ Run your program in RAMDISK.
- Copy your program directory to RAMDISK.
 cp ~/for/<prog-dir> /dev/shm/<user>/<prog-dir>
- Then run your program from RAMDISK. Running the program in the RAMDISK will decrease the load average and make your program runs faster.
- http://www.cyberciti.biz/tips/what-is-devshm-and-its-practical-usage.html