Problem with DRJIT_LIBLLVM_PATH environment variable

[dcamps]

Dear all,

I am running Ubuntu 22.04 on a VM running on virtualbox.

I run python and verify that the DRJIT_LIBLLVM_PATH environment variable is properly set by running:
>>> os.environ['DRJIT_LIBLLVM_PATH']
'/usr/lib/llvm-14/lib/libLLVM.so'

However, when I run import sionna, I get the following error:

2024-09-28 16:59:23.304419: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
jit_llvm_init(): your CPU does not support the fma instruction set, shutting down the LLVM backend...
Traceback (most recent call last):
File "/home/vagrant/sionna/lib/python3.10/site-packages/mitsuba/init.py", line 107, in getattribute
import('mitsuba.mitsuba' + variant + '_ext'),
File "/usr/lib/python3.10/importlib/init.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "", line 1050, in _gcd_import
File "", line 1027, in _find_and_load
File "", line 1006, in _find_and_load_unlocked
File "", line 674, in _load_unlocked
File "", line 571, in module_from_spec
File "", line 1176, in create_module
File "", line 241, in _call_with_frames_removed
ImportError: jit_init_thread_state(): the LLVM backend is inactive because the LLVM shared library ("libLLVM.so") could not be found! Set the DRJIT_LIBLLVM_PATH environment variable to specify its path.

I am not sure if the problem is the environment variable not being set for some reason, or if the error is related to the no support for FMA, which results in the shutdown of the LLVM backend.

Amy help is appreciated.

BR

Daniel

[merlinND]

Hello @dcamps,

Another user has reported better success using LLVM 16: #275 (comment)
Could you please try that version and report back?

Otherwise, there might be an issue related to the VM, the warning message "your CPU does not support the fma instruction set, shutting down the LLVM backend" is suspicious. Maybe there's a VirtualBox setting that allows you to enable FMA in the virtualized CPU?