GlastRelease Tag: rhel4_gcc34opt
System Tests Report Summary
No notable differences for comparison with GR v15r47p12gr10.
There are minor changes in a couple histograms that reflect small differences in the MC generation. This may reflect the processor used to run the test and its comparison version. (I'm doing some specific comparisons to check this more carefully. The system tests require rhel 4 queues, but there are processor differences within that set. This could explain the bizarre feature that the histograms are not just similar, but identical in number of entries and content for tests other than the 100 GeV case.) The distributions that are changing (ACD POCAs, CALTOWER) are those sensitive to changes internal event content, but do not indicate a large change or any apparent differences in the higher level behavior.
The last version that had careful scrutiny was GR v15r47p12gr06. There are more differences for that comparison, but the changes made since then are the likely cause - looking at that more closely now.