I was doing CTW runs with -d64 and tripped across a failure from nips.blockers.

CompileTheWorld (42) : symantec/tools/debug/RemoteValue
# To suppress the following error report, specify this argument
# after -XX: or in .hotspotrc:  SuppressErrorAt=/loopnode.hpp:573]
#
# An unexpected error has been detected by Java Runtime Environment:
#
#  Internal Error (/net/prt-solamd64-q1-4/PrtBuildDir/workspace/src/share/vm/opto/loopnode.hpp, 573), pid=21391, tid=11
#
# Java VM: Java HotSpot(TM) 64-Bit Server VM (20060905093858.nips.blockers-debug mixed mode)
#
# Error: assert(n != 0L,"Bad immediate dominator info.")
# An error report file with more information is saved as hs_err_pid21391.log
#
# If you would like to submit a bug report, please visit:
#   http://java.sun.com/webapps/bugreport/crash.jsp
#


You can reproduce it with this:

java -d64 -XX:+CompileTheWorld -XX:CompileTheWorldStartAt=42 -Xbootclasspath/p:/net/vmsqe.sfbay/export/backup/testbase/CompileTheWorld/jarfiles/scd10.jar
Posted Date : 2006-10-25 22:46:12.0

-XX:-ReassociateInvariants

reassociate_invariants used wrong method to register new
node which failed to add the new node to the member list of the loop
it was nested in. The loop was subsequently cloned into a
pre/main/post which didn't clone the new node correctly, causing the error.
Posted Date : 2006-10-26 19:10:44.0