From: Leandro Martínez (leandromartinez98_at_gmail.com)
Date: Fri May 18 2007 - 16:19:33 CDT
Hi Gengbin,
Using the "-memory os" option and 4 processors (the master and one node)
the simulation crashes as before:
------------- Processor 0 Exiting: Caught Signal ------------
Signal: segmentation violation
Suggestion: Try running with '++debug', or linking with '-memory paranoid'.
Stack Traceback:
[0] /lib/libc.so.6 [0x2ad41cf005c0]
[1] _ZN10Controller9threadRunEPS_+0 [0x5d80d0]
Fatal error on PE 0> segmentation violation
However, with this option and using only local processors (++local +p2),
the simulation starts running (we don't know if it is stable).
Exactly the same behavior is observed using the "-thread context" option.
Brian, we ran the "pgm" test for charm++, and it passed, both locally
and using all nodes. I don't have the log now of the
compilation, I will sent it to you when I compile it again.
Leandro.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:42 CST