[sectionBody]
Hello,
I am developing a complicated CFD code. During the development, I found that the codes generated by
Intel Fortran 10 and 11 produced different results, and the code from 11 seems obvious wrong. This puzzled
me for some time. Recently, I was able to reproduce the problem in the following simple code.
The expected output from the code is
1. 1. 1. 1. 1. 1. 1. 1.
-2. -2. -2. -2. -2. -2. -2. -2.
-3. -3. -3. -3. -3. -3. -3. -3.
However, the output with compiler 11 is
-1. -1. -1. -1. -1. -1. -1. -1.
-2. -2. -2. -2. -2. -2. -2. -2.
3. 3. 3. 3. 3. 3. 3. 3.
The problem occurs for win32 and x64 with or without optimization. Is it a bug of the compiler or a bug of our coding.
The compiler I have the problem is 11.0.3451.2005.
Thanks,
Haifeng
----------------------------------------------------------------------
module sd_data_mod
type mydattype
real,pointer :: fld(:,:,:,:)
end type mydattype
type(mydattype) :: mt
end module sd_data_mod
program main
use sd_data_mod
allocate(mt%fld(2,2,2,3))
mt%fld(:,:,:,1) = 1
mt%fld(:,:,:,2) = 2
mt%fld(:,:,:,3) = 3
call view(mt%fld(:,:,:,2:3),2)
write(*,'(100f4.0)') mt%fld(:,:,:,1)
write(*,'(100f4.0)') mt%fld(:,:,:,2)
write(*,'(100f4.0)') mt%fld(:,:,:,3)
! expected correct output are:
! 1. 1. 1. 1. 1. 1. 1. 1.
! -2. -2. -2. -2. -2. -2. -2. -2.
! -3. -3. -3. -3. -3. -3. -3. -3.
end program main
subroutine view(fld,ns)
use sd_data_mod
integer :: ns
real :: fld(2,2,2,ns)
fld = -fld
end subroutine view[/sectionBody]
-----------------------------------------------------------------------
