>>That's a good link.<<

Those measurements for programs forced to use just one core on a quad-core machine, you may also be interested in the measurements made when the programs are allowed to use all the cores -

http://shootout.alioth.debian.org/u32q/which-programming-lan...

and also measurements made on x64

http://shootout.alioth.debian.org/u64q/which-programming-lan...

>>The n-body and spectral-norm benchmarks were fastest in Fortran.<<

Maybe we should wonder if the Intel C++ compiler would produce programs comparable to those from the Intel Fortran compiler ;-)

>>Object orientated programming is used in a significant portion of all programs, it would be interesting how C/C++ performs in such benchmarks.<<

Maybe you're suggesting that OO C++ programs wouldn't perform as well as Java programs? But doesn't that highlight the advantage of being able to write programs in different styles?