A correction to a small typo on my previous email:
THE RESULTS ARE:
OS : Win 2000 sp4
PROCESSOR : Athlon 1800+ MP
Memory : 512 MB
Compilers :
CVF 6.6b = Compaq Visual Fortran 6.6b
g95 for windows
n= 10**8 10**9
--------------------------------------
CVF 6.6b
--------------------------------------
prog1: < 1 sec < 1 sec
prog2: 13 sec 2min 12sec
--------------------------------------
g95
--------------------------------------
prog1: 12 sec 1min 35sec (TYPO WAS HERE)
prog2: 12 sec 1min 35sec (TYPO WAS HERE)
----------------------------------------------------
Jose J. Canals-Cerda
Department of Economics
University of Colorado at Boulder
Campus Box 256
Boulder, Colorado 80309-0256
Phone: (303) 492-7869
Fax: (303) 492-8960
E-mail: [log in to unmask]
----------------------------------------------------
For art visit www.StudioArtworks.com
----------------------------------------------------
Quoting [log in to unmask]:
> Hello,
>
> I use Fortran on my research (stats/econometrics) and I was excited
> to learn that g95 was available for Windows OS. I decided to run a
> couple of simple programs just for fun.
>
>
> PROGRAM 1:
>
> program time
> implicit none
> real(8) :: x,y
> integer :: n,i
>
> n = 10**9
>
> x = 0.0d0; y = 0.0d0
>
> do i=1,n
> x = 1.0d0*i
> x = 1/x
> x = dexp(x)
> y = y + x
> end do
>
> write(*,*)i-1
>
> pause
>
> end program time
>
>
> PROGRAM 2:
>
> program time
> implicit none
> real(8) :: x,y
> integer :: n,i
>
> n = 10**9
>
> x = 0.0d0; y = 0.0d0
>
> do i=1,n
> x = 1.0d0*i
> x = 1/x
> x = dexp(x)
> y = y + x
> end do
>
> write(*,*)i-1,y
>
> pause
>
> end program time
>
>
> THE RESULTS ARE:
>
> OS : Win 2000 sp4
> PROCESSOR : Athlon 1800+ MP
> Memory : 512 MB
>
> Compilers :
> CVF 6.6b = Compaq Visual Fortran 6.6b
> g95 for windows
>
> n= 10**8 10**9
> --------------------------------------
> CVF 6.6b
> --------------------------------------
> prog1: < 1 sec < 1 sec
> prog2: 13 sec 2min 12sec
> --------------------------------------
> g95
> --------------------------------------
> prog1: 12 sec 12 sec
> prog2: 1min 35sec 1min 35sec
>
>
> I guess the anomaly with prog1 in CVF is due to some type of
> compiler optimization, otherwise it does not make sense.
>
> I have also performend the same analysis in a P3 1000 machine
> and g95 also outperforms CVF 6.6b (4min 23sec CVF and 2min 37sec for g95)
>
> I must say that I have been very happy using CVF for the past 4
> years. This test is clearly not enough to pass judgment on g95,
> but is a start.
>
> In any case, the work already done by the g95 group is
> truly remarkable. Thank you!
>
> NOTE:
>
> If anybody is interested, the exe programs (prog 2) can be found at:
>
> http://spot.colorado.edu/~canalsce/g95/
>
> you need to have g95 installed in your computer before you can run
> the g95time.exe program, win2000 may also be required (?).
>
> Jose.
> ----------------------------------------------------
> Jose J. Canals-Cerda
> Department of Economics
> University of Colorado at Boulder
> Campus Box 256
> Boulder, Colorado 80309-0256
> Phone: (303) 492-7869
> Fax: (303) 492-8960
> E-mail: [log in to unmask]
> ----------------------------------------------------
> For art visit www.StudioArtworks.com
> ----------------------------------------------------
>
>
>
> Quoting Toon Moene <[log in to unmask]>:
>
> > Vivek Rao wrote:
> >
> > > G95, a free Fortran 95 compiler developed primarily by Andy Vaught, now
> > > runs on Windows as well as Linux and Unix (including Mac OS X)
> > > platforms.
> >
> > I would take issue with "primarily developed by Andy Vaught". As I have
> > explained in previous comp.lang.fortran posts, a compiler the size of
> > g95/gfortran is not developed by a single person.
> >
> > Everyone has to build on previous work, see:
> >
> > http://groups.google.nl/groups?
> hl=nl&lr=&threadm=20041014102031.453e0dce.d.a.ham%
> 40citg.tudelft.nl&prev=/groups%3Fnum%3D25%26hl%3Dnl%26lr%3D%26group%
> 3Dcomp.lang.fortran%26start%3D200
> >
> > Cheers,
> >
> > --
> > Toon Moene - e-mail: [log in to unmask] - phone: +31 346 214290
> > Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
> > Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
> > A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/
> >
>
|