This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Some (small) c++ compilation profiling data (oprofile)


John Levon wrote:-

> Again, not really scientific but :

Thanks, this is interesting, but I don't claim to understand it.

> (This is subject to frailties in bfd_find_nearest_line(),
> and the inaccuracy of delayed profiling interrupt delivery ..)
> 
> /* time count, time %age, mem ref count, mem reg %age */
> 
>     141 /* ht_lookup 149193 3.769% 27240 2.39% */

I looked on the oprofile site but couldn't see this format of
output, nor an explanation of what it meant.

However, I guess that the number is a cycle count (or something
equivalent).  But I don't understand why there are 2 numbers
and 2 percentages.

>     163   /* 37996 0.9599% 9582 0.8409% */
>     164   for (;;)

I don't understand this either - how can a for statement
be expensive?  Looking at the numbers shows this is not
a total for the loop, but is in addition to everything else.

Neil.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]