SGI F77 COMPILER BASIC TEST
                                99/02/01


OBJECTIVE

To measure a simple vector/parallel REAL*8 program speed using basic 
compiler optimization options.


DESCRIPTION

Machine: uranus.ijs.si (Indy R5000 SC/150, IP22: 1 x 150 MHz, 64 MB RAM).
Program: Matrix multiplication using nested DO loops.
Matrix dimensions: ND = 800 (reservation).
Timing: dtime (tempd.f)
Program details:
    P25:    I-loop inside (vector style), no directives:

      PROGRAM P25
      ...
      IMPLICIT    REAL*8  (A-H,O-Z)
      PARAMETER   (ND = 800, NIN = 5, NOUT = 6)
      DIMENSION   A(ND,ND), B(ND,ND), C(ND,ND)
      ...
      DO  26  J = 1,N
          DO  24  K = 1,N
              DO  22  I = 1,N
                  C(I,J) = C(I,J) + A(I,K) * B(K,J)
   22         CONTINUE
   24     CONTINUE
   26 CONTINUE
      ...


RESULTS


Table I.  Additional optimization options. ND = 400, N = 400 (128 M
operations).
--------------------------------------------------------------------------
Program   Compiler call                               CPU time      MFLOPS
--------------------------------------------------------------------------
p25       f77 -O3                                          8           16
              -O3 -Ofast=ip22                              2           64
	      -O3 -r5000 -mips3 -n32                       2.3         56
	      -O3 -r5000 -mips4 -n32                       1.5         85
	      -O3 -r5000 -mips4 -n32 -LNO:ou=4             1.5         85
	      -O3 -r5000 -mips4 -n32 -LNO:ou=6             1.5         85
--------------------------------------------------------------------------



Table II.  As Table I, but ND = 800, N = 800 (1024 M operations).
--------------------------------------------------------------------------
Program   Compiler call                               CPU time      MFLOPS
--------------------------------------------------------------------------
p25       f77 -O3                                         71           14
              -O3 -Ofast=ip22                             15           68
              -Ofast=ip22                                 15           68
              -Ofast=ip22                             
	          -LNO:ou=6:cs1=32k:cs2=512k -TENV:X=4    16           64
	      -O3 -r5000 -mips3 -n32                      21           49
	      -O3 -r5000 -mips4 -n32                      17           60
	      -O3 -r5000 -mips4 -n32 
	          -LNO:ou=4                               14           73
	          -LNO:ou=6                               13           79
	          -LNO:ou=6 -TENV:X=4                     13           79
	          -LNO:ou=6:cs1=32k:cs2=512k -TENV:X=4    13           79
	          -LNO:ou=8                               16           64
--------------------------------------------------------------------------



Table III.  As Table II, but p25v.f, calling DGEMM.
--------------------------------------------------------------------------
Program   Compiler call                               CPU time      MFLOPS
--------------------------------------------------------------------------
p25v
--------------------------------------------------------------------------