Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[ptp-user] TSC development

Hi all,

I’m still working on adapting Eclipse PTP to work with “K” computer.

I have some questions/problems:

[1] In my TSC file I have submit-batch block, and in match _expression_ of stdout-parser I get the correct jobId — I can see it in debug logs. I also have get-job-status block. Job status (while the job is running) is picked up correctly. 
  Nevertheless, in Progress window I can see two progress bars: Running Application and Lauching myAppName. Both of them at 100%. The problem is that they never go away even after the job has finished. 
  I attach my TSC file.

PNG image


[2] I can run profiling job with TAU. I get profile files, but they don’t get added to profile database. 
I set up a new database for profiles:
/applications/tau/tau/apple/bin > ./perfdmf_configure
Configuration file NOT found...
a new configuration file will be created.

Welcome to the configuration program for PerfDMF.
This program will prompt you for some information necessary to ensure
the desired behavior for the PerfDMF tools.


You will now be prompted for new values, if desired.  The current or default
values for each prompt are shown in parenthesis.
To accept the current/default value, just press Enter/Return.

Please enter the name of this configuration.
():imac
Please enter the database vendor (oracle, postgresql, mysql, db2, derby or h2).
(h2):
Please enter the JDBC jar file.
(/Applications/TAU/tau/apple/lib/h2.jar):
Please enter the JDBC Driver name.
(org.h2.Driver):
Please enter the path to the database directory.
(/Users/peterbryzgalov/.ParaProf/imac):
Please enter the database username.
():eclipse
Store the database password in CLEAR TEXT in your configuration file? (y/n):y
Please enter the database password:
Please enter the PerfDMF schema file.
(/Applications/TAU/tau/etc/taudb.sql):

Writing configuration file: /Users/---/.ParaProf/perfdmf.cfg.imac

Now testing your database connection.

Database created, command: jdbc:h2:/Users/---/.ParaProf/imac/perfdmf;AUTO_SERVER=TRUE;create=true

Uploading Schema: /Applications/TAU/tau/etc/taudb.sql
Found /Applications/TAU/tau/etc/taudb.sql  ... Loading
Successfully uploaded schema

Database connection successful.
Configuration complete.

I ran install_plugins.sh and can see the DB in Performance Data Manager with name “default”. In profiling configuration window I can select my DB for storing profiles. But the profiles get never stored there. 

PNG image


I suspect it is because job finish is not detected. And thus my first question.  

[3] Running monitoring on “K” computer is a problem: processing XML files by perl scripts takes much longer then a minute. Is there a way to speed up this process? 
    K computer has about 83 000 nodes. Running commands to get a job list and a node list doesn’t take much time. But after node list and job list are ready, combining XML files by perl scripts takes a few minutes.  

[4] After my TSC more or less works, I would like to share it, so that other users of Fujitsu computers can use it. What is the right way to share TSC and server-side (supercomputer-side) scripts?
  
Best regards,
Peter


Attachment: PJM-Generic.xml
Description: XML document



Back to the top