Fortran求大数阶乘
2014-03-10 15:42:10   来源：Fcode研讨团队   评论：0 点击：

program www_fcode_cn
    implicit none
    integer(2) nLbound, nUbound
    real t1,t2
    call cpu_time(t1)
    nLbound = 1
    nUbound = 32767
    call factorial( nLbound, nUbound )
    call cpu_time(t2)
    print*, 'CPU_Time:',t2-t1
    pause
 end program www_fcode_cn

 subroutine factorial( nLbound, nUbound )
 Use, Intrinsic :: ISO_FORTRAN_ENV, only : OUTPUT_UNIT
 implicit none
 integer(2),intent(in) :: nLbound, nUbound
 integer(8),parameter  :: pmax = 10**13, len = 10300
 integer(8) ires(len), nL, nU
 integer(8) i, j, m, n
 character ch*24

 ires=0; ires(1)=1
 nL = 1; nU = 1
 do i = nLbound, nUbound
     ires(nL:nU) = ires(nL:nU) * i
     n = 0
   do j = nL, nU
     if( ires(j)/=0 ) then
       m = ires(j) / pmax
       ires(j+1) = ires(j+1) + m
       ires(j) = ires(j) - pmax*m
       n = 1
     else if( n==0 ) then
       nL = j
     end if
   end do
   if( ires(nU+1) > 0 ) nU = nU + 1
 end do

 ! 输出结果：n!
 open(output_unit,file='1.txt') !输出到文件

 write(ch,*) ires(nU)
 ch = adjustl(ch)
 m = len_trim(ch)  !首元素中整数值的位数
 n = (nU-1)*13+m-1 !指数

 !科学计数法
 write(OUTPUT_UNIT,'(a)',advance='no') ch(1:1)//'.'//ch(2:m)//'e+'
 write(ch,*) n
 ch = adjustl(ch)
 write(OUTPUT_UNIT,'(a)') trim(ch)
 !全部数字
 if( nU > 1 ) then
   write(OUTPUT_UNIT,'(i<m>)',advance='no') ires(nU)
   write(OUTPUT_UNIT,'(<nu-1>i13.13)') ires(nU-1:1:-1)
 else
   write(OUTPUT_UNIT,'(i<m>)') ires(nU)
 end if 
 close(OUTPUT_UNIT)
 end subroutine factorial