Re: Release of Fluka2011.2.7: respin of Fluka2011.2

... we'll do a 4.5.x version in the next days and post it to the web site.
However it will not solve the possible miscompilation of user routines
for those having gfortran 4.4.x on their machines and it is anyway a
bandaid solution for the 4.6.x-4.5.x incompatibiltities since we cannot
go on forever in building versions on different subversions of the same
compiler.

                     Ciao
                    Alfredo

+----------------------------------------------------------------------+
| Alfredo Ferrari || Tel.: +41.22.76.76119 |
| CERN-EN/STI || Fax.: +41.22.76.69474 |
| 1211 Geneva 23 || e-mail: Alfredo.Ferrari_at_cern.ch |
| Switzerland || |
+----------------------------------------------------------------------+

On Thu, 17 Nov 2011, Vassili Maroussov wrote:

> Dear Alfredo,
>
> It would be very kind of FLUKA team to offer the option to download the
> gfortran 4.5.x compatible x86_64 libraries. As it was correctly mentioned,
> there is no other solution for users in a case if executable should be
> compiled with user routines (which in their turn rely on 3rd party libraries,
> etc.)
>
> Best regards,
>
> Vassili
>
> On 11/16/2011 06:23 PM, Alfredo Ferrari wrote:
>> Hi all
>>
>> the issue with the gfortran version (and the gfortran version only), is
>> that it has been built on gfortran 4.6.1 (we upgraded recently our
>> machines) and the gfortran developers apparently didn't find any
>> better joke than changing in between gfortran 4.5.x and gfortran 4.6.x
>> the names of some internal compiler procedures, in particular those that
>> in the past were called
>>
>> _gfortran_transfer_xxx
>>
>> are now called
>>
>> _gfortran_transfer_xxx_write
>>
>> so obviously fluka compiled on 4.6.x it does not link on 4.5.x or 4.4.x
>> (and possibly viceversa, to be checked)
>>
>> Without commenting on the sadistic attitude of the gfortran developers,
>> if what I described above holds true to a deeper investigation, there is
>> no simple solution. One possibility is for the user to upgrade their
>> gfortran compiler to 4.6.x (very uncomfortable to say the least), or for
>> us to build a version on some remaining machines still with gfortran 4.5.1
>> and provide two versions if required (maybe the one compiled on 4.5.x
>> links on 4.6.x). It is clear that this kind of incompatible changes
>> between compiler sub-versions if confirmed are a major pain...
>>
>> Meawhile for those experiencing these problems the g77 version works as
>> usual.
>>
>> I would like also to warn all users running with gfortran 4.4.x
>> (eg those on Scientific Linux) that we have evidence that gfortran 4.4.x
>> can miscompile Fluka routines. It is not an issue if you only link the
>> standard executable from a library compiled with 4.5.x or higher, it can
>> be a major problem if you have user routines, which obviously you compile
>> locally with gfortran 4.4.x. We had one group which clearly demonstrated
>> this problem recently with their source routine.
>> Again, we are powerless versus compiler incompatibilities/bugs, we can
>> only suggest to upgrade the compiler in this case.
>>
>> Ciao
>> Alfredo
>>
>> +----------------------------------------------------------------------+
>> | Alfredo Ferrari || Tel.: +41.22.76.76119 |
>> | CERN-EN/STI || Fax.: +41.22.76.69474 |
>> | 1211 Geneva 23 || e-mail: Alfredo.Ferrari_at_cern.ch |
>> | Switzerland || |
>> +----------------------------------------------------------------------+
>>
>> On Wed, 16 Nov 2011, Joseph Comfort wrote:
>>
>>> Hi Alfredo,
>>>
>>> I also had trouble installing Fluka2011.2.7. From an earlier habit, even
>>> the
>>> command
>>> flutil/lfluka -m fluka
>>> (nearly the same as 'make flukahp') does not work. The error messages are
>>> essentially the same as previously reported. Note that ...2.6 was fine.
>>> I
>>> had expected a quick message about a correction. The file I downloaded in
>>> not useable, and has been deleted.
>>>
>>> Joe
>>>
>
Received on Thu Nov 17 2011 - 14:54:05 CET