Does Sparse Solver support ilp64 format?

[ino]

I use PARDISO routines in MKL Library. It works fine when it deals with a modest number of equations.
However, when the number of equations gets large, I've encountered a linking error as below.
I use ifort with "-i8" option. I'm dealing with array with dimension slightly less than one billion.
According to the user reference of MKL library, the sparse linear solver
is supposed to support the ilp64 format which handles arrays with dimension as large as two billions.
How can I get rid of these linking errors?

/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(diag.o)(.text+0x770): In function `L_diag_pardiso__76__par_region2':
: relocation truncated to fit: R_X86_64_PC32 ::___kmpv_zerodiag_pardiso(void)
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(diag.o)(.text+0x8d6): In function `L_diag_pardiso__76__par_region2':
: relocation truncated to fit: R_X86_64_PC32 ___kmpv_zerodiag_pardiso__2
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_diag.o)(.text+0x837): In function `L_c_diag_pardiso__69__par_region2':
: relocation truncated to fit: R_X86_64_PC32 ::___kmpv_zeroc_diag_pardiso(void)
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_diag.o)(.text+0x9a0): In function `L_c_diag_pardiso__69__par_region2':
: relocation truncated to fit: R_X86_64_PC32 ___kmpv_zeroc_diag_pardiso__2
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(do_all_fc.o)(.text+0x3ac3): In function `L_do_all_pardiso_fc__963__par_region0':
: relocation truncated to fit: R_X86_64_PC32 ::___kmpv_zerodo_all_pardiso_fc(void)
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(do_all_fc.o)(.text+0x3fbf): In function `L_do_all_pardiso_fc__963__par_region0':
: relocation truncated to fit: R_X86_64_PC32 ___kmpv_zerodo_all_pardiso_fc__3
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_pre_cgs.o)(.text+0xed0): In function `c_pre_cgs_pardiso_':
: relocation truncated to fit: R_X86_64_PC32 ::___kmpv_zeroc_pre_cgs_pardiso(void)
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_cgs.o)(.text+0x92f): In function `L_c_cgs_a__95__par_region0':
: relocation truncated to fit: R_X86_64_PC32 ::___kmpv_zeroc_cgs_a(void)
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_cgs.o)(.text+0x1117): In function `c_cgs_b_':
: relocation truncated to fit: R_X86_64_PC32 ___kmpv_zeroc_cgs_b__2
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_amuxy_res.o)(.text+0x437): In function `c_amuxy_res_pardiso_':
: relocation truncated to fit: R_X86_64_PC32 .bss
/opt/intel/mkl/10.0.3.020/lib/em64t/libmkl_solver_ilp64.a(c_amuxy_res.o)(.text+0x7ed): In function `c_amuxy_res_pardiso_':
: additional relocation overflows omitted from the output

[Sergey_P_Intel2]

Hello!

There is too little information in your post, namely, it is clear only that mkl10.0.3 is used. In order to investigate this problem we need more information, namelyhardware/OS configuration, versions of compiler and linker, linkage string, environment variables and, if possible test case to reproduce the situation.

Could you provide us with more details about the problem?It would be better to post QuADrequest with some attached files via https://premier.intel.com.

All the best,

Sergey