[petsc-dev] [petsc-maint #129715] __float128 complex

[Jose E. Roman]

f2cblaslapack is more or less prepared for this case already (it is not much tested though, we will make some tests for the w* subroutines tomorrow).

What is not ready is PETSc itself. Some definitions are missing. For instance, PetscScalar is undefined, so one should add the following to petscmath.h:106:

#elif defined(PETSC_USE_REAL___FLOAT128)
typedef __complex128 PetscScalar;

#define PetscRealPart(a)      crealq(a)
#define PetscImaginaryPart(a) cimagq(a)
#define PetscAbsScalar(a)     cabsq(a)
#define PetscConj(a)          conjq(a)
#define PetscSqrtScalar(a)    csqrtq(a)
#define PetscPowScalar(a,b)   cpowq(a,b)
#define PetscExpScalar(a)     cexpq(a)
#define PetscLogScalar(a)     clogq(a)
#define PetscSinScalar(a)     csinq(a)
#define PetscCosScalar(a)     ccosq(a)

Similarly, MPIU_SCALAR should be defined. Maybe other as well.

Jose


El 22/08/2012, a las 15:04, Barry Smith escribió:

> 
>    The f2cblaslapack is generated by the script bin/maint/toclapack.sh   from BLAS/LAPACK source code. 
> 
>    Jose and his students fixed up the support for __float128;    presumably it could be extended to __complex128 also with some tweaking.
> 
> 
>     Barry
> 
> 
> 
> On Aug 22, 2012, at 6:52 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
> 
>> GNU __float128 does not work with complex. There is __complex128 (why they
>> chose to implement this way, I do not understand), but we have not written
>> that support yet. You can "watch" this issue to get notified when it's done
>> 
>> https://bitbucket.org/petsc/petsc-dev/issue/6/support-for-complex-with-128-bit-precision
>> 
>> Barry, is just doing text substitution to f2cblaslapack the right approach
>> there?
>> 
>> On Wed, Aug 22, 2012 at 3:38 AM, Alexander Grayver
>> <agrayver at gfz-potsdam.de>wrote:
>> 
>>> Hello petsc team,
>>> 
>>> I've tried to compile petsc-3.3-p2 with __float128 precision and complex
>>> scalar type. The configure stage went through, but during make there are
>>> errors.
>>> See attached make.log
>>> 
>>> Any ideas?
>>> 
>>> --
>>> Regards,
>>> Alexander
>>> 
>>> 
>>> 
>> 
>