1 |
double precision function ddot(n,dx,incx,dy,incy) |
2 |
c |
3 |
c forms the dot product of two vectors. |
4 |
c uses unrolled loops for increments equal to one. |
5 |
c jack dongarra, linpack, 3/11/78. |
6 |
c modified 12/3/93, array(1) declarations changed to array(*) |
7 |
c |
8 |
double precision dx(*),dy(*),dtemp |
9 |
integer i,incx,incy,ix,iy,m,mp1,n |
10 |
c |
11 |
ddot = 0.0d0 |
12 |
dtemp = 0.0d0 |
13 |
if(n.le.0)return |
14 |
if(incx.eq.1.and.incy.eq.1)go to 20 |
15 |
c |
16 |
c code for unequal increments or equal increments |
17 |
c not equal to 1 |
18 |
c |
19 |
ix = 1 |
20 |
iy = 1 |
21 |
if(incx.lt.0)ix = (-n+1)*incx + 1 |
22 |
if(incy.lt.0)iy = (-n+1)*incy + 1 |
23 |
do 10 i = 1,n |
24 |
dtemp = dtemp + dx(ix)*dy(iy) |
25 |
ix = ix + incx |
26 |
iy = iy + incy |
27 |
10 continue |
28 |
ddot = dtemp |
29 |
return |
30 |
c |
31 |
c code for both increments equal to 1 |
32 |
c |
33 |
c |
34 |
c clean-up loop |
35 |
c |
36 |
20 m = mod(n,5) |
37 |
if( m .eq. 0 ) go to 40 |
38 |
do 30 i = 1,m |
39 |
dtemp = dtemp + dx(i)*dy(i) |
40 |
30 continue |
41 |
if( n .lt. 5 ) go to 60 |
42 |
40 mp1 = m + 1 |
43 |
do 50 i = mp1,n,5 |
44 |
dtemp = dtemp + dx(i)*dy(i) + dx(i + 1)*dy(i + 1) + |
45 |
* dx(i + 2)*dy(i + 2) + dx(i + 3)*dy(i + 3) + dx(i + 4)*dy(i + 4) |
46 |
50 continue |
47 |
60 ddot = dtemp |
48 |
return |
49 |
end |