I have seen this message with ifort for as long as I have been living
with ARPACK. If you check the dsarpack code, you'll see that the
compiler is wrong. The "jump" is perfectly fine.
Istvan
Quoting Scott Brozell <sbrozell.rci.rutgers.edu>:
> Hi,
>
> On Mon, Mar 01, 2010 at 04:19:36PM -0500, Ben Roberts wrote:
>>
>> I've been trying to debug something unrelated, so I built Amber with
>> ifort -check, and I get this response:
>>
>> dsarpack.f(199): warning #6371: A jump into a block from outside the
>> block may have occurred. [12]
>> go to (12,18) label
>> -------------^
>> dsarpack.f(199): warning #6371: A jump into a block from outside the
>> block may have occurred. [18]
>> go to (12,18) label
>> ----------------^
>>
>> Is arpack something we should be playing around with, and if so, does
>> this message represent a problem to be fixed?
>
> No, this like other long living libraries (blas, ...) generally shouldnt
> be cleaned up by us. There are several perspectives here, but the simplest
> is that they have been tested extensively, and most compilers are
> not designed to decrypt their spaghetti code.
>
> Scott
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Mar 02 2010 - 07:30:04 PST
