On quarta-feira, 22 de agosto de 2012 13.21.03, [email protected] wrote: > Wow… do you get the same results with Qt 4? I did quite a bit to optimize > contention performance in Qt 4. I do know that mutex contention performance > has regressed in Qt 5 on the Mac, I wasn't sure about on Linux, though.
4.8 results, on the same machine, same build options, same everything:
(second best result out of 9 runs, to exclude outliers)
RESULT : tst_QMutex::uncontendedQMutex():
62.2731147 CPU ticks per iteration
465.803163 task-clock # 0.999 CPUs utilized
1,556,403,915 cycles # 3.341 GHz
1,728,088,651 instructions # 1.11 insns per cycle
271 raw_syscalls:sys_enter # 0.582 K/sec
0.466074468 seconds time elapsed
RESULT : tst_QMutex::contendedQMutex():"no msleep, 1 mutex":
2,235,753.069 CPU ticks per iteration
6331.041527 task-clock # 3.758 CPUs utilized
20,161,652,971 cycles # 3.185 GHz
15,429,162,599 instructions # 0.77 insns per cycle
9,022,998 raw_syscalls:sys_enter # 1.425 M/sec
1.684687915 seconds time elapsed
RESULT : tst_QMutex::contendedQMutex():"no msleep, 2 mutexes":
4,368,748.283 CPU ticks per iteration
10941.737969 task-clock # 3.786 CPUs utilized
34,877,133,530 cycles # 3.188 GHz
26,077,481,434 instructions # 0.75 insns per cycle
14,471,550 raw_syscalls:sys_enter # 1.323 M/sec
2.889920428 seconds time elapsed
--
Thiago Macieira - thiago.macieira (AT) intel.com
Software Architect - Intel Open Source Technology Center
Intel Sweden AB - Registration Number: 556189-6027
Knarrarnäsgatan 15, 164 40 Kista, Stockholm, Sweden
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Development mailing list [email protected] http://lists.qt-project.org/mailman/listinfo/development
