On Tue, 31 Mar 2009, Szak�ts Viktor wrote: > Attached results. > Both with -gc3, one with r10748 (stack calls), and one > with r10756 + HB_USER_CFLAGS=-6r. ST and MT, > two runs with each combination. > [ The previously posted quick MT results seems wrong, now > with repeated and saved tests it's about 64s (not 75s). Still > not stellar, but better. ] > I don't want to draw any conclusions, -6r mode causes > a slight overall slowdown in ST mode and a minor speedup > in MT mode. Slowdown in ST mode can be accounted for > T030 and T031, the rest of tests became faster.
Thank you. In my tests passing parameters by register: [ total application time: ]...................35.83 [ total real time: ]..........................35.87 and by stack: [ total application time: ]...................39.90 [ total real time: ]..........................39.92 It's single thread mode. I'll try to make some tests on different machines with real windows. What is the size difference between both binaries? Can you upload them to some public host? best regards, Przemek _______________________________________________ Harbour mailing list [email protected] http://lists.harbour-project.org/mailman/listinfo/harbour
