angela
it looks to me that the erfcfun call on line 3397 of ewald.f is in the
middle of an #ifdef segment and will be compiled only if EXACT_SHORT is
in your set of compile time definitions in config.h---otherwise a table
lookup is used in its place. Note that at startup time this table is built
using erfcfun----look in fill_erf_table at line 1906
regards
Tom D
On Tue, 13 Sep 2005, Angela Liu wrote:
> Dear Caoch,
>
> I have fixed the problem. Here is what I did, and you
> can tell me if there might be potential problems or
> not:
>
> I borrowed the erfcfun.f in directory /src/sander/,
> removed the first two "#include" lines to remove the
> dependency on "dprec.h", and commented out all the
> "call derfcfun()" lines in the /src/gibbs/ewald.f
> file. Then I added erfcfun.f and erfcfun.o to the
> /gibbs/Makefile at corresponding places. The
> compilation went smoothly, and now I have an
> executable "gibbs".
>
> The only puzzle left is that on line 3397 of file
> /gibbs/ewald.f, the "call erfcfun(x, erfc)" function
> call returns a new value for "erfc", but it is never
> used, unlike the other function calls of "erfcfun()"
> in this file. I wonder if this is done for
> convenience or not (so as to avoid writing an extra
> function?).
>
> Thanks very much for your helpful suggestions!
>
> Best regards,
> Angela
>
> --- caoch.cherry.bio.titech.ac.jp wrote:
>
>> Dear angela:
>>
>> I don't know which OS, Machine and fortran
>> compiler your are using.
>> anyway, if you insist on using derfcfun your should
>> get a sutiable
>> erfcfun.c file for your machine. then put it into
>> sys.a (
>> $AMBERHOME/src/lib).
>>
>> regards
>>
>> caoch
>>> Dear Caoch,
>>>
>>> After searching for the "call derfcfun(x, erfcc)"
>>> lines, I found that there are many references to
>> erfcc
>>> and x that are modified by this function. This
>> means
>>> that many other variable assignments are also
>>> affected. Wouldn't this mean that those lines
>> need to
>>> be motified as well? Is there a simple way to
>> download
>>> the required library file onto my computer?
>>>
>>> Thank you!
>>>
>>> Angela
>>>
>>>
>>>> --- caoch.cherry.bio.titech.ac.jp wrote:
>>>>
>>>>> dear angela:
>>>>>
>>>>>
>>>>>> 3) How can I get hold of the source code of
>>>>> "gibbs" in
>>>>>> Amber 7 or 6? I tried "make" in the
>>>>>> $AMBERHOME/src/gibbs/ directory, but there
>> were
>>>>>> problems with the Ewald routine (RE erfcfun)
>>>> that
>>>>>> caused the "make" to fail. I wasn't sure how
>>>> this
>>>>> can
>>>>>> be corrected.
>>>>>
>>>
>>>>> for some reasons it seems that some OS don't
>>>>> contain erfcfun liberary. i have checked
>>>>> the gibbs source codes. With my opinion
>> derfcfun
>>>>> shouldn't effect the caculation
>>>>> accuracy.So just remove "call derfcfun" line
>> and
>>>>> compile the source codes.
>>>>>
>>>>> good luck
>>>>> caoch
>>>>>>
>>>>>> Thank you very much for your time,
>> suggestions,
>>>>> and
>>>>>> help!
>>>>>>
>>>>>> Best wishes,
>>>>>> Angela Liu
>>>>>>
>>>>>>
>>>>>>
>>>>
>> __________________________________________________
>>>>>> Do You Yahoo!?
>>>>>> Tired of spam? Yahoo! Mail has the best
>> spam
>>>>> protection around
>>>>>> http://www.rosswalker.co.uk/adsense_alternatives/
>>>>>>
>>>>>
>>>>
>>>
>>
> -----------------------------------------------------------------------
>>>>>> The AMBER Mail Reflector
>>>>>> To post, send mail to amber.scripps.edu
>>>>>> To unsubscribe, send "unsubscribe amber" to
>>>>> majordomo.scripps.edu
>>>>>>
>>>>>
>>>>> --
>>>>> caoch.cherry.bio.titech.ac.jp
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> __________________________________
>>>> Yahoo! Mail - PC Magazine Editors' Choice 2005
>>>> http://www.rosswalker.co.uk/adsense_alternatives/
>>>>
>>>
>>
> -----------------------------------------------------------------------
>>>> The AMBER Mail Reflector
>>>> To post, send mail to amber.scripps.edu
>>>> To unsubscribe, send "unsubscribe amber" to
>>>> majordomo.scripps.edu
>>>>
>>>
>>>
>>>
>>>
>>> __________________________________
>>> Yahoo! Mail - PC Magazine Editors' Choice 2005
>>> http://www.rosswalker.co.uk/adsense_alternatives/
>>>
>>
> -----------------------------------------------------------------------
>>> The AMBER Mail Reflector
>>> To post, send mail to amber.scripps.edu
>>> To unsubscribe, send "unsubscribe amber" to
>> majordomo.scripps.edu
>>>
>>
>> --
>> caoch.cherry.bio.titech.ac.jp
>>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://www.rosswalker.co.uk/adsense_alternatives/
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Tue Sep 13 2005 - 16:53:00 PDT
