Using fgmres to solve large-scale linear equation systems

[houjy]

 I have used ILP64 successfully.

But I have another question,When I use fgmres to solve large systems of linear equations, some values in the ipar array are equal to -5589219598436180572 

This is my code example. I would like to try solving a system of linear equations with a scale of 69335390 or above. Thank you very much.

[houjy]

May I ask what this means? The environment I am using is the Visual Studio compiler for Windows

[houjy]

Is there any other solution or elimination method

[Chao_Y_Intel]

MKL_INT is used in the code. It need to add "MKL_ILP64" micro to define this type as 64 bit integer. 

To compiler it with MSVC compiler:

1)Set the Target Application to x64 (64-bit Code):

2)Add the MKL_ILP64 Micro Definition:

Go to Project > Properties > C/C++ > Command Line.

In the Additional Options box, add the following micro definition: /DMKL_ILP64

3)Add ILP64 Libraries to the Linker:

 Project > Properties window > Linker > Input. Add the following ILP64 libraries in  

  mkl_intel_ilp64.lib mkl_intel_thread.lib mkl_core.lib libiomp5md.lib

4)Also need to set the include/ Library library path: 

 Project>Properties > VC++ Directories> Include Directories, Add MKL Include Path: xxx\mkl\latest\include

Project>Properties >Linker >General >Additional Library Directories, Add MKL library Path:  xxx\mkl\latest\lib