[3dem] GPU-ERROR - Relion-3.0-beta
tnakane at mrc-lmb.cam.ac.uk
Mon Feb 4 23:45:29 PST 2019
Which version of nvcc did you use for compilation?
Do you have the right version of CUDA runtime in LD_LIBRARY_PATH?
On 2019/02/05 5:20, Dieter Blaas wrote:
> Dear all,
> I did a test install of the latest version of Relion-3.0-beta on a
> relatively potent workstation (CentOS Linux release 7.5.1804) and
> everything runs fine on all 64 CPUs. However, when running e.g. Class2D
> using 2 or 4 GPUs (set to 5 MPIs/4 Threads and 3 MPIs/2 Threads,
> respectively) I receive the error below immediately upon starting the
> run. The 4 GPUs are 1080 Ti (11 GB). To the best of my knowledge none of
> the possibilities below applies. I am not very familiar with compilation
> and cannot exclude that I did something wrong. However, there was no
> error during the compilation with cmake3 followed by make and I
> previously installed this version of relion on two other machines
> without any issue!
> Thanks for hints, bw Dieter
> ERROR: unknown error in
> at line 128 (error-code 30)
> line 67
> === Backtrace ===
> /usr/lib64/libc.so.6(__libc_start_main+0xf5) [0x7fc01f87a3d5]
> A GPU-function failed to execute.
> If this occured at the start of a run, you might have GPUs which
> are incompatible with either the data or your installation of relion.
> If you
> -> INSTALLED RELION YOURSELF: if you e.g. specified -DCUDA_ARCH=50
> and are trying ot run on a compute 3.5 GPU (-DCUDA_ARCH=3.5),
> this may happen.
> -> HAVE MULTIPLE GPUS OF DIFFERNT VERSIONS: relion needs GPUS with
> at least compute 3.5. You may be trying to use a GPU older than
> this. If you have multiple generations, try specifying --gpu <X>
> with X=0. Then try X=1 in a new run, and so on. The numbering of
> GPUs may not be obvious from the driver or intuition. For a list
> of GPU compute generations, see
> -> ARE USING DOUBLE-PRECISION GPU CODE: relion was been written so
> as to not require this, and may thus have unforeseen requirements
> when run in this mode. If you think it is nonetheless necessary,
> please consult the developers with this error.
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
More information about the 3dem