Dear hoomd developers
I have been trying to run simulations of a pretty large macromolecule using Hoomd. I am using a standard model (FENE+LJ+some bending potential). I have managed to install Hoomd version 2.9.6 (the reason being I was given a script and I did not want to update it to the 3.0.0 version) with CUDA 11.4 on a system with a Tesla T4
Everything goes well with the installation but, once I ran my script (that I tested elsewhere), I get the following error
** starting run **
**ERROR**: __global__ function call is not configured before /hoomd/Autotuner.cc:129
Traceback (most recent call last):
File "denpol_test.py", line 236, in <module>
hoomd.run (to_run)
File "/home/ubuntu/myvenv/lib/python3.6/site-packages/hoomd/__init__.py", line 201, in run
context.current.system.run(int(tsteps), callback_period, callback, limit_hours, int(limit_multiple));
RuntimeError: CUDA Error
**ERROR**: invalid resource handle before /hoomd/GPUArray.h:160
terminate called after throwing an instance of 'std::runtime_error'
what(): CUDA Error
[hoomdblue:04479] *** Process received signal ***
[hoomdblue:04479] Signal: Aborted (6)
[hoomdblue:04479] Signal code: (-6)
(and afterwards a bunch of memory errors). I have tried to look for the errors, without much success. It looks like it is indeed a memory error, but I do not understand how can it be. It's strange also because the same happens even if I choose --mode=cpu.
Can you help me?
Thank you in advance
Emanuele Locatelli