Re: GNAT compiler switches and optimization

[Gautier <gautier@fakeaddress.nil>]

tkrauss:
> I'm a bit stuck trying to figure out how to coax more performance
> out of some Ada code.  I suspect there is something simple (like
> compiler switches) but I'm missing it.  As an example I'm using
> a simple matrix multiply and comparing it to similar code in
> Fortran.  Unfortunately the Ada code takes 3-4 times as long.

One thing is already that you start the timer at the wrong place.
You should start it after filling your array with random numbers.
In the present state you compare the random generator, then your matrix 
multiplication. It is possible that Ada.Numerics.Float_Random.Random takes 
significantly more time due to Ada's quality requirements in the random 
generation.
On the other hand, switches you can try are:
  -O2 instead of -O3, -funroll-loops (usually good) -ffast-math,
  for both Ada and Fortran
  -gnatn, for Ada
Since you are measuring real time and not CPU time, you might want also to 
take a bit larger matrices in order to have disk swaps and rests of 
initializations effects statistically small.

HTH, Gautier
______________________________________________________________
Ada programming -- http://www.mysunrise.ch/users/gdm/gsoft.htm

NB: For a direct answer, e-mail address on the Web site!!