ERROR: no CUDA-capable device is detected

[Dominique]

Hello,

I get this error when trying to run Gctf:

ERROR: no CUDA-capable device is detected in /home/EM_Packages/relion2.0/src/ctffind_runner.cpp at line 189 (error-code 38) 

nvidia-smi gives me:

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

Any suggestions would be appreciated.

Best,

Dominique

[Dominique]

Update: Running via the command line on a single micrograph gave me this:

Error in function Gctf_get_parameters(), file ctf52.cu on line 5374: Error: Failed to find any proper GPU device!

! Exit now!

Gpuinfo gives me:

Simple GPU info query.

You have 1 nVidia GPGPU.

DeviceID Name Version Memory(Mb)

Failed to get GPU info of Device #0

I am Using EM-packages 4.2GPU

was this the correct AMI to use?

[dominiqueg...@gmail.com]

 lspci -vnn | grep VGA 

gives me only (below) and not a nvidia VGA compatible controller  : 

00:02.0 VGA compatible controller [0300]: Cirrus Logic GD 5446 [1013:00b8] (prog-if 00 [VGA controller])

[Michael Cianfrocco]

Hi Dominque, 

DId you get these errors from the cryoem-cloud-tools submission through the RELION GUI, or are these problems associated with running jobs directly on an AWS instance? 

Mike

[Dominique]

Hello,

I got this error on the relion GUI:

ERROR: no CUDA-capable device is detected in /home/EM_Packages/relion2.0/src/ctffind_runner.cpp at line 189 (error-code 38) 

The other errors I got running jobs on the command line of the instance.

[Michael Cianfrocco]

Hi Dominique,

I just submitted a job successfully using Gctf and the RELION GUI submission to AWS. Can you tell me more about the job information that you provided to run this? 

-Region (e.g. us-west-2)

-RELION input options 

-Gctf options in RELION GUI

Thanks,

Mike