Difference between revisions of "Too Large Events"

From New IAC Wiki
Jump to navigation Jump to search
Line 88: Line 88:
 
and the event number+1  
 
and the event number+1  
 
<pre>
 
<pre>
eventnum=eventnum entry=272)
+
eventnum=eventnum entry=272
 
</pre>
 
</pre>
  
which causes the error.  Since the random number sequence is set in the gcard, we can edit the LUND file and insert new values for the Moller electron (all 0).  The random number sequence hopefully continues without more too large DC hit numbers.
+
which causes the error.  Since the random number sequence is set in the gcard, we can edit the LUND file and insert new values for the Moller electron (all 0).   
 +
<pre>
 +
273event #*3lines per event=819 line number
 +
</pre>
 +
 
 +
 
 +
The random number sequence hopefully continues without more too large DC hit numbers.

Revision as of 18:49, 11 January 2018

For some evio files, the events are too large. The errors are:

The lines below might hint at the cause of the crash.
If they do not help you then please submit a bug report at
http://root.cern.ch/bugs. Please post the ENTIRE stack trace
from above as an attachment in addition to anything else
that might help us fixing this issue.
===========================================================
#5  std::_Rb_tree<unsigned int, std::pair<unsigned int const, bankIndex_t>, std::_Select1st<std::pair<unsigned int const, bankIndex_t> >, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, bankIndex_t> > >::_M_erase (this=this
entry=0x7f6162e57e40, __x=0xc038b9e9502d69cf) at /usr/include/c++/4.8/bits/stl_tree.h:1125
#6  0x00000000004426e5 in clear (this=0x7f6162e57e40) at /usr/include/c++/4.8/bits/stl_tree.h:860
#7  clear (this=0x7f6162e57e40) at /usr/include/c++/4.8/bits/stl_map.h:789
#8  EvioBankIndex::initBankIndex (this=0x7f6162e57e40, buffer=0x7f6161f15a40) at src/reader/EvioBankIndex.cc:24
#9  0x000000000043cdc3 in EvioDataEvent::init (this=this
entry=0x7f6161f15a40, ptr=<optimized out>, len=<optimized out>) at src/reader/EvioDataEvent.cc:136
#10 0x000000000043c5aa in EvioFileReader::readEvent (this=this
entry=0x7f6161d8f010, eventnum=eventnum
entry=496) at src/reader/EvioFileReader.cc:114
#11 0x0000000000409787 in main (argc=<optimized out>, argv=<optimized out>) at src/convertor/evio2root.cc:851
===========================================================

Running an eviodump on the file we find

*********************** EVENT # 497  ***********************

        +------------------------------------------------+------------+------------+
        |                                            bank|       nrows|       ncols|
        +------------------------------------------------+------------+------------+
        |                                        DC::dgtz|        5568|          11|
        |                                        DC::true|        5568|          24|
        |                                   GenPart::true|           2|           7|
        +------------------------------------------------+------------+------------+


The gcard is limited in what it records for HITS.

 <option name="BANK_VERBOSITY" value="1"/>
        <option name="HIT_VERBOSITY" value="1"/>
        <option name="LOG_VERBOSITY" value="2"/>
        <option name="G4TRACK_VERBOSITY" value="2"/>
        <option name="G4P_VERBOSITY" value="2"/>
        <option name="GEN_VERBOSITY" value="2"/>
        <option name="SAVE_ALL_MOTHERS" value="2"/>
        <option name="ENERGY_CUT" value="-1"/>
        <option name="RECORD_PASSBY" value="1"/>


The variable in the evio2root files for such events are unsigned integer (uint32_t) of values 0 to 4294967295. Each value is stored in binary bits, i.e. 2^0=1, 2^1=2 -> 3 decimal = 0011. Since it is unsigned this limits the value stored to only be positive values.

The value above has the hex value of entry=0x7f6162e57e40 76f6162e57e40 hex=2092788898692672 dec

2092788898692672-4294967295=2092784603725377 which is over the size of the 32bit variable. We could edit the entire evio read program structure for a 64bit variable or ignore just that specific Theta-Momentum pair. The output of the Shield in may have more DC interactions, but the effect of the shield still blocks electrons regardless if it is on or off. The plots shown earlier and the paper show there isn't much difference. Simulations will be done with the FT OFF.

For Shield IN/OFF -1xTorus Field we still get a too large event.

The lines below might hint at the cause of the crash.
If they do not help you then please submit a bug report at
http://root.cern.ch/bugs. Please post the ENTIRE stack trace
from above as an attachment in addition to anything else
that might help us fixing this issue.
===========================================================
#5  std::_Rb_tree<unsigned int, std::pair<unsigned int const, bankIndex_t>, std::_Select1st<std::pair<unsigned int const, bankIndex_t> >, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, bankIndex_t> > >::_M_erase (this=this
entry=0x7fc8a4ecae40, __x=0x40b0c0168b89c126) at /usr/include/c++/4.8/bits/stl_tree.h:1125
#6  0x0000000000442725 in clear (this=0x7fc8a4ecae40) at /usr/include/c++/4.8/bits/stl_tree.h:860
#7  clear (this=0x7fc8a4ecae40) at /usr/include/c++/4.8/bits/stl_map.h:789
#8  EvioBankIndex::initBankIndex (this=0x7fc8a4ecae40, buffer=0x7fc8a3f88a40) at src/reader/EvioBankIndex.cc:24
#9  0x000000000043ce03 in EvioDataEvent::init (this=this
entry=0x7fc8a3f88a40, ptr=<optimized out>, len=<optimized out>) at src/reader/EvioDataEvent.cc:136
#10 0x000000000043c5ea in EvioFileReader::readEvent (this=this
entry=0x7fc8a3e02010, eventnum=eventnum
entry=272) at src/reader/EvioFileReader.cc:114
#11 0x0000000000409797 in main (argc=<optimized out>, argv=<optimized out>) at src/convertor/evio2root.cc:851
===========================================================


We know the file

OPENED FILE : LH2_0Sol_-1Tor_11GeV_IsotropicPhi_ShieldIn_081/LH2_0Sol_-1Tor_11GeV_IsotropicPhi_ShieldIn_081.evio


and the event number+1

eventnum=eventnum entry=272

which causes the error. Since the random number sequence is set in the gcard, we can edit the LUND file and insert new values for the Moller electron (all 0).

273event #*3lines per event=819 line number


The random number sequence hopefully continues without more too large DC hit numbers.