New compiler sanity check

[Richard Felkins_2]

cat h.f

real*4 y,yy
yy=1.10
y=0.0
do j=1,1000000
y=y+yy
enddo
write(*,*)' single precision y = ',y
stop
end


18% ./h
single precision y = 1110920.
19%

New FORTRAN 5.5 installation fails check after
installation. The answer should be 1100000 and
not 1110920 the actual result.

Anyone have an idea while this is exhibiting
what appears to be an underflow problem.

Thanks,

Rich.