[Pw_forum] QE-GPU (Submitting calculations)

Phanikumar Pentyala phani12.chem at gmail.com
Mon Mar 6 12:47:34 CET 2017


Hello QE users

I have single server attached with two gpu accelerators (details mentioned
below).  Currently I am in stage of submitting calculations, So I want to
ask some questions regarding that

1. I want to control usage of GPU accelerators, with usage of some GPU
cores or one GPU device at a time or two GPUs at a time when submitting of
calculation. Is it possible to control that?

2. *mpirun -np xx pw-gpu.x <input> output.log &* , This is the general
command we used for submitting calculations. Is there anything I had to add
to this command?


*Details of server:*

linux: *centOS*

Server:

*FUJITSU Server PRIMERGY RX2540 M2*
Total no. of processors:
* 40 *

GPU card:
*NVIDIA Tesla K40m (2 number)*

CUDA:

*CUDA Runtime Version = 8.0*
QE version:


* QE-GPU version 5.4*
Thanks in advance


Regards
Phanikumar
Research Scholar
Chemical Engineering Department
Indian institute of technology Kharagpur (IITK)
India
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20170306/63806eb0/attachment.html>


More information about the users mailing list