Compilation errors – dkt2d, settlingCube3d

[sthavishtha]

Dear all

I am using the latest version of olb-v1.3-1. Unfortunately, both the example codes – dkt2d and settlingCube3d throw the same compilation errors as shown below. I have also tried make cleanbuild, make clean and make, which still yield the same compilation errors.

Compile dkt2d.cpp
g++ -DOLB_PRECOMPILED -O3 -Wall -march=native -mtune=native         -std=c++14  -I/cluster/home/sbhopala/olb/src -I/cluster/home/sbhopala/olb/src/ -I/cluster/home/sbhopala/olb/src/external -I/cluster/home/sbhopala/olb/src/external/zlib -c dkt2d.cpp -o /cluster/scratch/sbhopala/run_ictslbm/dkt2d/dkt2d.o
In file included from /cluster/home/sbhopala/olb/src/functors/lattice/functors2D.hh:31:0,
                 from /cluster/home/sbhopala/olb/src/functors/functors2D.hh:30,
                 from /cluster/home/sbhopala/olb/src/olb2D.hh:5,
                 from dkt2d.cpp:23:
/cluster/home/sbhopala/olb/src/functors/lattice/blockLatticeLocalF2D.hh: In instantiation of ‘bool olb::BlockLatticePorousMomentumLossForce2D<T, DESCRIPTOR>::operator()(T*, const int*) [with T = double; DESCRIPTOR = olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR>]’:
dkt2d.cpp:268:1:   required from here
/cluster/home/sbhopala/olb/src/functors/lattice/blockLatticeLocalF2D.hh:1050:13: error: no matching function for call to ‘olb::Cell<double, olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR> >::setField(double [2])’
             this->_blockLattice.get(iX, iY).template setField<descriptors::VELOCITY_NUMERATOR>(reset_to_zero);
             ^
In file included from /cluster/home/sbhopala/olb/src/core/blockLatticeStructure2D.h:31:0,
                 from /cluster/home/sbhopala/olb/src/boundary/boundaryCondition2D.h:31,
                 from /cluster/home/sbhopala/olb/src/boundary/boundary2D.h:28,
                 from /cluster/home/sbhopala/olb/src/olb2D.h:1,
                 from dkt2d.cpp:22:
/cluster/home/sbhopala/olb/src/core/cell.h:153:3: note: candidate: template<class FIELD, class X> olb::utilities::meta::enable_if_t<(X:: size<FIELD>() > 1), void> olb::Cell<T, DESCRIPTOR>::setField(const olb::Vector<T, DESCRIPTOR:: size<FIELD>()>&) [with FIELD = FIELD; X = X; T = double; DESCRIPTOR = olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR>]
   setField(const Vector<T,DESCRIPTOR::template size<FIELD>()>& field)
   ^
/cluster/home/sbhopala/olb/src/core/cell.h:153:3: note:   template argument deduction/substitution failed:
/cluster/home/sbhopala/olb/src/core/cell.h: In substitution of ‘template<class FIELD, class X> olb::utilities::meta::enable_if_t<(X:: size<FIELD>() > 1), void> olb::Cell<T, DESCRIPTOR>::setField(const olb::Vector<T, DESCRIPTOR:: size<FIELD>()>&) [with FIELD = olb::descriptors::VELOCITY_NUMERATOR; X = <missing>]’:
/cluster/home/sbhopala/olb/src/functors/lattice/blockLatticeLocalF2D.hh:1050:13:   required from ‘bool olb::BlockLatticePorousMomentumLossForce2D<T, DESCRIPTOR>::operator()(T*, const int*) [with T = double; DESCRIPTOR = olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR>]’
dkt2d.cpp:268:1:   required from here
/cluster/home/sbhopala/olb/src/core/cell.h:153:3: error: expression ‘size<olb::descriptors::VELOCITY_NUMERATOR>’ is not a constant-expression
/cluster/home/sbhopala/olb/src/core/cell.h:153:3: note: in template argument for type ‘unsigned int’ 
/cluster/home/sbhopala/olb/src/functors/lattice/blockLatticeLocalF2D.hh: In instantiation of ‘bool olb::BlockLatticePorousMomentumLossForce2D<T, DESCRIPTOR>::operator()(T*, const int*) [with T = double; DESCRIPTOR = olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR>]’:
dkt2d.cpp:268:1:   required from here
/cluster/home/sbhopala/olb/src/core/cell.h:164:3: note: candidate: template<class FIELD, class X> olb::utilities::meta::enable_if_t<(X:: size<FIELD>() == 1), void> olb::Cell<T, DESCRIPTOR>::setField(T) [with FIELD = FIELD; X = X; T = double; DESCRIPTOR = olb::descriptors::D2Q9<olb::descriptors::POROSITY, olb::descriptors::VELOCITY_NUMERATOR, olb::descriptors::VELOCITY_DENOMINATOR>]
   setField(T value)
   ^
/cluster/home/sbhopala/olb/src/core/cell.h:164:3: note:   template argument deduction/substitution failed:
In file included from /cluster/home/sbhopala/olb/src/functors/lattice/functors2D.hh:31:0,
                 from /cluster/home/sbhopala/olb/src/functors/functors2D.hh:30,
                 from /cluster/home/sbhopala/olb/src/olb2D.hh:5,
                 from dkt2d.cpp:23:
/cluster/home/sbhopala/olb/src/functors/lattice/blockLatticeLocalF2D.hh:1050:13: note:   cannot convert ‘reset_to_zero’ (type ‘double [2]’) to type ‘double’
             this->_blockLattice.get(iX, iY).template setField<descriptors::VELOCITY_NUMERATOR>(reset_to_zero);
             ^
make: *** [/cluster/scratch/sbhopala/run_ictslbm/dkt2d/dkt2d.o] Error 1

I made appropriate changes in the ROOT and SRC of definitions.mk. However, I don’t believe this to be the source of mistake as the parallel and sequential versions of cavity2d work.

Regards
Sthavishtha

[mathias]

Please compile using the generic option in confik.mk according to:

#BUILDTYPE := precompiled
BUILDTYPE := generic

Dont forget to “make cleanbuild”.

Best
mathias

[sthavishtha]

Dear Dr. Mathias

I forgot to mention in my earlier post. I did try with a generic BUILDTYPE as well. Yet, I get the same error.

Regards
Sthavishtha

[mathias]

Ok, then resinstall everything without any changes and report if that clean version works in generic mode. It should work. From that clean starting point, apply step by step your wanted chanches and leave a meassage again in that forum if something doesnt work. Then, also state exactly which compiler you are using.

[sthavishtha]

Dear Dr. Mathias

Thanks for the comments.

The error arises due to the compiler version. I now use GCC 6.2.0 and it works. Using an old compiler version (GCC 5.2.0) earlier was enough to compile cavity2d, cylinder2d, cylinder3d problems. So, I was unable to identify the issue earlier with the examples available in the particles/ folder.

Regards
Sthavishtha

[Author]

Posts