Hi,
I've ran the n-body benchmark code in my machine (Debian Linux 3.2.0-2-amd64 #1 SMP x86_64 GNU/Linux -- Intel Xeon X5482 3.20GHz) using the latest Go tree (hg pul/hg up @ the time of this message) and got some really good results (see below)!
I'm sure my Gfortran compilation is wrong since Go is almost twice as faster than the Fortran code... Also, the C++ and C codes are faster than the ADA one (opposite to what is shown in
shootout.alioth.debian.org).
Does anyone have further experience on this? (for convenience, I'm attaching some files from
shootout.alioth.debian.org alongside my Makefile)
Cheers.
Dorival
./run.bash
running ADA code
-0.169075164
-0.169059907
real 0m10.994s
user 0m10.993s
sys 0m0.000s
running CPP code
-0.169075164
-0.169059907
real 0m8.333s
user 0m8.333s
sys 0m0.000s
running C code
-0.169075164
-0.169059907
real 0m9.628s
user 0m9.617s
sys 0m0.008s
running Go code
-0.169075164
-0.169059907
real 0m12.388s
user 0m12.385s
sys 0m0.000s
running Fortran code
-0.169075164
-0.169059907
real 0m21.222s
user 0m21.213s
sys 0m0.008s
running Python (2.7) code
-0.169075164
-0.169059907
real 14m20.470s
user 14m19.814s
sys 0m0.680s
running Python (3.2) code
-0.169075164
-0.169059907
real 13m35.734s
user 13m35.727s
sys 0m0.032s
should output:
-0.169075164
-0.169059907