0

I am doing a relaxation calculation using gpu in a system of 800,000 atoms. Sometimes, I get the following error. Can you please tell me the cause and how to fix it?

Cuda driver error 700 in call at file 'geryon/nvd_timer.h' in line 76.
application called MPI_Abort(MPI_COMM_WORLD, -1) - process 0
talonmies
  • 70,661
  • 34
  • 192
  • 269
ev ance
  • 3
  • 1
  • That is a runtime error probably caused by an addressing error within the LAMMPS GPU code itself, in all likelihood, or something like the code taking too long on your GPU or running out of memory. You should report it to the LAMMPS developers. It isn't a CUDA related error and you are very unlikely to get help here for a problem like that – talonmies Jul 06 '21 at 04:08

1 Answers1

0

In short, it has been solved by changing different versions of CUDA and ARCH code. If you want to use 2080, you must choose cuda10.2 with -DGPU_ARCH=sm_70. II have 3090 and 2080 on the HPC. I thought it was downward compatible but it is not. I googled it that someone says that 2080 does not support cuda11.4.

Matt Andruff
  • 4,974
  • 1
  • 5
  • 21
Roy_Kid
  • 41
  • 5