[parsec-users] counting locks of blackscholes
fahim_yazdan at yahoo.com
Thu Feb 12 05:59:31 EST 2009
As an experience, I wrote a pintool to count synchronization primitives such as lucks, I think it works correctly, and most of its results are the same as table.2 in parsec technical report, but some of them are different. For example, about blackscholes, the number of locks is reported 17 by my program in the following address(it is also reported by ManualExample/proccount of pin examples), but in the table 2 it is 0.
pthread_mutex_lock libpthread.so.0 b65b79d0 17
My supervisor wants to know the reason of this difference. Would you please make me sure about this? I mean whether this function must be counted for this package or by a reseon must not.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the parsec-users