Hi,
I have an OEM model compiled for a 32-bit (x86) architecture, but our OPAL-RT target uses a 64-bit (x64) architecture.
Is there a way to import and run this model in HYPERSIM 2026, or does it need to be recompiled for x64?
Best regards,
Juan
Hi Juan,
The short answer is no, you cannot directly run a 32-bit (x86) precompiled model on a 64-bit (x64) OPAL-RT target. The model must be recompiled for 64-bit (x86_64) architecture.
Why recompilation is required:
Target Architecture & Runtime: Modern OPAL-RT real-time targets run a 64-bit operating system (OPAL-RT Linux 64-bit), and HYPERSIM executes simulation tasks as native 64-bit processes on the target CPU.
Binary Incompatibility: In Linux and real-time execution environments, a 64-bit process cannot link against or dynamically load 32-bit binary shared libraries (.so, .a, or .dll) due to differences in memory addressing and ABI.
Recommended Next Steps:
If you have the source code (.c, .cpp, .h):
You can import the source files directly into HYPERSIM using the UDM / UCM (User Defined Component) workflow or C-code block interfaces. HYPERSIM will automatically compile and link the model using the 64-bit target toolchain during the build/load process.
If the OEM model is a black-box precompiled binary:
You will need to contact the OEM/vendor and request a 64-bit Linux (x86_64) compiled build of the library.
If the model uses a standardized interface (such as IEEE/CIGRE DLL or FMI/FMU), ask the vendor for the Linux x64 binary/export.