Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.

Exception thrown when calling DGESVD

spainchaud
New Contributor I
482 Views

I am using VS2022 and oneAPI 2024.2. I have a Fortran DLL that I have linked to MKL using /Qmkl:parallel. I call the DLL from a C++ program. When I reach the first call to MKL (DGESVD), I get this error message.

Exception thrown at 0x00007FF838AFB699 (KernelBase.dll) in Toolbox2dGuid.exe: 0xC06D007E: Module not found (parameters: 0x00000015FFF937C0).
Unhandled exception at 0x00007FF838AFB699 (KernelBase.dll) in Toolbox2dGuid.exe: 0xC06D007E: Module not found (parameters: 0x00000015FFF937C0).

I can see in the stack that KernalBase.dll is called from mkl_intel_thread.2.dll.

This code has been run in the past as part of a console program, and in a static library, so I know there is no problem there. I suspect a problem with my setup/installation because I have to copy mkl_intel_thread.2.dll to my executable directory for the C++ application to run. The link instructions in the MKL developer seem to indicate I just have to select parallel or sequential. I have actually tried both with the same lack of success.

 

 

0 Kudos
1 Solution
spainchaud
New Contributor I
423 Views

I resolved the problem by turning on the MKL libraries in the C++ project in addition to the Fortran DLL. I no longer need to have mkl_intel_thread.2.dll placed in the executable directory. The libraries are switched on at Project > Properties > Configuration Properties > Intel Libraries for oneAPI > Intel oneAPI Math Kernal Library > Use oneMKL.

View solution in original post

0 Kudos
1 Reply
spainchaud
New Contributor I
424 Views

I resolved the problem by turning on the MKL libraries in the C++ project in addition to the Fortran DLL. I no longer need to have mkl_intel_thread.2.dll placed in the executable directory. The libraries are switched on at Project > Properties > Configuration Properties > Intel Libraries for oneAPI > Intel oneAPI Math Kernal Library > Use oneMKL.

0 Kudos
Reply