RFC: fp printing speedup patch

Jerry Quinn jlquinn@optonline.net
Fri Nov 14 06:32:00 GMT 2003


Paolo Carlini writes:
 > Hi Jerry. I'm just back and still trying to catch up...
 > 
 > Anyway, I have already printed the patch and was hoping to return
 > to you tomorrow or at most on monday.
 > 
 > Some things I can tell you immediately:
[snip] 
 > 
 >     2- Some more numbers about performance: for sure it's a big
 >        improvement (thanks again) but which is still the gap wrt v2
 >        for some interesting examples (big number of digits/small
 >        number of digits)

On numbers that are handled by the bigint code, floatconv is not too
fast.  On the following test, I get:

real    0m9.812s
user    0m9.790s
sys     0m0.010s

as opposed to ~2.8 seconds when the fast path is used for relatively
small amounts of precision.  For reference, I get from gcc-2.95:

real    0m8.392s
user    0m8.360s
sys     0m0.000s

and gcc-3.3

real    0m25.976s
user    0m25.910s
sys     0m0.010s

Jerry



#include <iostream>
#include <iomanip>

int main()
{
  for (int i = 0; i < 3000000; i++) {
    std::cout << std::setprecision(30)
	 << (double)i + 9.09923452563465256292345256346525629234525634652562e+18 << '\n';
  }
  return 0;
}



More information about the Libstdc++ mailing list