I've got the problem. I spent too much time on a function which is deeply inside the loop. most of the caculation is not necessary, as most of them just yield repeating result. I'll try modifying the code. Thanks everybody! On 4/21/05, Georg Steffers <georg@xxxxxxxxxxxx> wrote: > As far as I understand profiling you must not read the time results as > the values you got without profiling. Profiling is there to find > inefficiencies in your code, such as unneccessary loops etc. and thus > the times are only to compare the different calls. Additionally it is > very likely that gprof called you app several times, to get a better > comparison. (That is only how i understand profiling, dont beat me if i > am wrong) > > You can see that your program spend the most time for the logarithmic > calculation, followed by the power caculation. > > In my opinion your problem is that you did not use fast-math, as far as > i know vc uses a math lib with similar optimizations by default. > > so as Ishwar Rattan mentioned, if speed is important and vc6 is faster > that use vc6. > > greets Georg > > Am 21.04.2005 08:17:12 schrieb(en) CyberSnoopy: > > I've tried gprof > > this is the result > > Flat profile: > > > > Each sample counts as 0.01 seconds.(what does this mean?) > > % cumulative self self total > > time seconds seconds calls s/call s/call name > > 32.23 7.56 7.56 __ieee754_log > --snip-- > > 0.02 23.45 0.01 > --snip-- > > I'm wondering is the total time as much as 23.46 seconds? I'm > > comfused... > > >
