Hi, I run bodytrack (a parsec program) on 16 cpus with 16 threads, one thread per cpu. cpu0 send much more barriers than other processors.
system.cpu00.commit.membars 51427 # Number of memory barriers committed system.cpu01.commit.membars 1 # Number of memory barriers committed system.cpu02.commit.membars 1 # Number of memory barriers committed system.cpu03.commit.membars 1 # Number of memory barriers committed system.cpu04.commit.membars 1 # Number of memory barriers committed system.cpu05.commit.membars 1 # Number of memory barriers committed system.cpu06.commit.membars 1 # Number of memory barriers committed system.cpu07.commit.membars 1 # Number of memory barriers committed system.cpu08.commit.membars 1 # Number of memory barriers committed system.cpu09.commit.membars 1 # Number of memory barriers committed system.cpu10.commit.membars 1 # Number of memory barriers committed system.cpu11.commit.membars 1 # Number of memory barriers committed system.cpu12.commit.membars 1 # Number of memory barriers committed system.cpu13.commit.membars 1 # Number of memory barriers committed system.cpu14.commit.membars 1 # Number of memory barriers committed system.cpu15.commit.membars 1 # Number of memory barriers committed
_______________________________________________ gem5-users mailing list gem5-users@gem5.org http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
