Difference between revisions of "Dan's Log"

From Nuclear Physics Group Documentation Pages
Jump to navigationJump to search
Line 188: Line 188:
 
<pre>
 
<pre>
 
GEMC is not running on Pumpkin.  On attempting to start it, get  
 
GEMC is not running on Pumpkin.  On attempting to start it, get  
 +
 
*************************************************************
 
*************************************************************
 
  Geant4 version Name: geant4-09-01    (14-December-2007)
 
  Geant4 version Name: geant4-09-01    (14-December-2007)

Revision as of 19:35, 29 July 2008

Progress and Task Log for Daniel

Tasks

  • Finalize/clean-up build instructions.
  • DONE - Figure-out how to add elements to GUI
  • DONE - Steal Kyle's rather spiffy wiki formatting
  • Get rid of red lines
  • Add way to change material of target in GUI
  • Add way to change particle beam in GUI
  • Display commands issued by GUI
  • Need to be able to turn on magnetic fields
  • DONE - Need command interface
  • Ability to list/change variables
  • Need list/glossary of GEANT4 commands (linked from JLAB wiki)
  • ^Plus, need our changes
  • Document XML
  • Make simple, limited XML description
  • Model single chip
  • Make nice macros
  • Are SVT5 and SVT6 micromegas? Is Sim12 model still accurate?
  • Are material descriptions incomplete? If so, should we fill-out?
  • Get sanitized version of Sim12 source
  • Understand how SVT works
  • Write summary of how SVT works

Misc. Notes

  • nTrigger is the number of times the beam is fired.
  • Sim12/Material_HDDS.xml says that XML spec is at http://www.gluex.org/hdds, but get a 404. Maybe $HDDSROOT/HDDS-1_0.xsd is what was there?

Log

Dan's Old Logs

June 2008

Tuesday 3 13:00-16:30


Thursday 5 13:00-16:30


Monday 9 13:00-16:30

Right.  Must get back on top of things.  Also, record time more consistently.
Today I'm going to make a tutorial for newbies.
...
Spent a bunch of time trying to get my environment to work.  When I source my pumpkin
environment script, my shell turns unresponsive.  Doing the same thing worked last
week.
Steps to reproduce:
[dan@benfranklin ~]$ ssh -X pumpkin.unh.edu
dan@pumpkin.unh.edu's password:
Last login: Mon Jun  9 15:45:07 2008 from benfranklin.unh.edu
[dan@pumpkin ~]$source /net/data/pumpkin2/CLAS12/RHEL5_x64/env.sh
[dan@pumpkin ~]$ ls
man gcc
blah
anything

*sigh*
CTRL-C doesn't help, nor does CTRL-D.
It's the LD_LIBRARY_PATH line.
Either I am being very stupid, or something is broken.  Probably the former.
It's dying when attempting to access /net/data/taro/maurik/root/root
There is no /net/data/taro
Automount broken?
Or Taro?
Well, Taro has moved.  Changed environment script, now it works.  Bah.  Going
home, hopefully I'll get something done there....

Tuesday 10 13:00-16:00

Pumpkin was having issues, so we all went down to the computer room.
Parity was replaced.
Worked with Kyle and Sarah on getting useful things out of ROOT.
I felt like I was a little bit lost, plus the temporary lack of a computer meant that I couldn't
contribute much.  Pretty unsatisfying day.  Need to go bone-up on ROOT, GEMC, and Physics.

Wednesday 11 13:00-16:00

Most of our time today was wasted with software issues.  We changed ROOT installs due to 
/net/data/taro/maurik disappearing, and this had annoying effects.  Losing Parity and the net
access issue didn't help.  
Ran a revised 50000 event test, but could not analyze because evio2root was linked against
old ROOT.  So we pointed it at a crippled install which was still there, which let it build, but the
binary caused segmentation violations.  Left annoyed and feeling like nothing was working.
Bah!

Remotely, after hours, I built a fresh ROOT, in /net/data/pumpkin2/CLAS12/RHEL5_64/ROOT
It works, as does the evio2root I built against it.  I also installed ROOT on Alice, and set-up
the computers in our new room.  I cannot seem to compile our ROOT programs on Alice, though.


Thursday 12

14:00-16:30

Tuesday 17

13:00-14:30

Wednesday 18

14:15-17:15
Details to follow...

Thursday 19

13:15-16:30
Installed EVIO2ROOT in $CLAS12ROOT, compiled for RH5_x64.
Sarah showed us a cool logging program called elog.  I installed it and ran it with the default configuration,
but then found that I could not kill the process.  This could be a problem as we don't have administrative 
privileges, even on the workstations.  I haven't spent much time thinking about how it is possible that I
couldn't kill it.  I've never figured-out how a user can create a SUID binary (if it was).  Anyway, configuring 
that took a back seat to other things.  Also, we should probably run it on our workstations rather than
on pumpkin.
I built and installed Valgrind to find the memory leak(s) in gemc.  I don't really know what I'm doing yet, but
it sure seemed like Valgrind said that there was not much leaking happening.  It sure slows things down
though.  Of course, I also don't know anything about manually allocating/deallocating memory, as I have 
previously only used languages with garbage collection.  So at least I should learn something.
Kyle and I messed with doing some larger runs for purposes of tutorials, but it didn't come to much.
I am still behind him in terms of familiarity with root, and desperately want to catch up.  Perhaps I will be
able to this weekend.
Oh, in order to compile gemc with debugging symbols and no crazy optimizations, I couldn't find any
easier way than to manually edit the SConstruct file, adding a CCFLAGS and CPPFLAGS option to the
environment.  # ADD BETTER EXPLANATION#

Tuesday 24

13:30-15:45
Worked on getting elog to work, and valgrind.  Met with Sarah and Kyle.

Thursday 26

13:00-17:00
Got ELOG to work, though I cannot seem to access it from anywhere but a local machine.  I suspect a packet filter. Most of the time I spent on that, plus some time doing time sheets and working on my wiiki guide and tidying things in the CLAS12 directories.

PAID UNTIL HERE


July 2008

Tuesday 1 13:00-16:30


Thursday 10 11:00-14:45

Kyle and I messed around with computers in the office before giving up on them.  We went down into the Physics Lounge, and installed Ubuntu
on them on the sly.  Things were actually pretty productive, with Kyle helping me catch up on my ROOT knowledge.  Eventually talked to
Sarah a little and clarified what I should be doing on the tutorial.  In short:  Focus on new students trying to use GEMC.


Friday 11 08:00-10:30

Got here early to work on things before the meeting.  Things not working, eventually decided UNH network is down.  GAH!
Tried to get things done despite it.  Naturally, things start working five minutes before our meeting.  *sigh*


Monday 14 13:00-19:00

GEMC is not running on Pumpkin.  On attempting to start it, get 

*************************************************************
 Geant4 version Name: geant4-09-01    (14-December-2007)
                      Copyright : Geant4 Collaboration
                      Reference : NIM A 506 (2003), 250-303
                            WWW : http://cern.ch/geant4
*************************************************************

  >>> gemc Init: >>  Retrieving gemc Detector Map...

  >>> gemc MySQL: >>  Connecting to database "clas12_geometry" as "clasuser". 
terminate called after throwing an instance of 'std::out_of_range'
  what():  basic_string::substr
Aborted
[dan@pumpkin ~]$ 

Was working last week, and I didn't change anything.  Either a pumkpin issue, or database.  My money is on the
database.  Will investigate later.

Working with ROOT.
Found something very useful:  To spit all of the attributes of an object on a canvas to the terminal, 
right-click on it and click on "dump".  Alternatively, clock on "Inspect"

GAH!  How can I change the scale of the color palette axis?  I found a reference to fwmax.  
Tweaked Kyle's phi_z.C
Made edep_strip.C
Sent an email to Maurizio asking for an explanation regarding the bands in phi vs. z graph.
It seems as though you are not supposed to change the palette scale yourself, or at least they make it
terribly difficult.  It seems always to be set automatically, and is highly sensitive to the values in one's
histogram.  This coupled with the fact that ours is different from Maurizio's suggests that we are doing
something differently from him.

I learned a lot about ROOT today.
Oh, use  gStyle->SetNumberContours(255); for a nicer gradient in your color palette.
Also, I don't think Maurizio cut the data for phi=0 from his phi_z graph, as "0" appears as the lower bound.
If he had set it higher, 0 would not have appeared, based on my experience.  But hopefully he will say.
Oh, some database info:  http://clasweb.jlab.org/wiki/index.php/12/10/2007_hands_on

Edep vs strip

Tuesday 15 13:00-14:45, 15:45-17:00

Met with Maurik and Sarah.
Should see if GEMC crashes on the farm.  Login to login.jlab.org, then ssh to ifarml1, remembering to
forward X.  But where is it?  Where can I run it from?  It is not in /apps.  Ask Maurizio.

Thursday 22

Wrapped-up first draft of new user guide.
Put-up copy of Mauruzio's First SVT Results page, started filling with our tests.  With source code!


Monday 28 14:00-17:00


Touching-up ROOT programs and uploading them to UNH SVT Results page.
My p vs phi does not look like Maurizio's, but seems to make sense.  Why does his "curve" at low momentums?
Asked him.  My theta vs z also looks different, will investigate next time.
Tried accessing elog on Sarah's workstation with her instructions, but couldn't get it to work.

Tuesday 29 11:30-14:30

Maurizio said that he used a magnetic field on the p vs phi graphs, explaining the discrepancy.  Did he
on others?  I will generate a new data set and re-run our ROOT programs, and see what happens.
Reading through the GEMC code, trying to trace variables through.