[Pw_forum] QE GPU acceleration pool performance

[Ivan Girotto]

Dear 子逸 張,

first of all I would like to remind you that there is a special mailing 
list for such discussion.

GPUs are nothing magic but simply massively parallel devices usually 
attached to a CPU multi-core board.
The complexity of accelerated systems requires a considerable 
technological knowledge.
If lacking of this fundamental component I strongly recommend to keep 
considering the CPU version of the distribution, only.
However, with good balance between CPU and GPU, high-end modern 
technology and a significant computational workload, decent speed up was 
registered while using GPU on top of the CPU platform.

Kind regards,

Ivan

On 09/04/2014 11:19, 子逸 張 wrote:
> Hello,
>
> I have some questions about the GPU version of QE.
>
> In my testcases,  GPU version of QE didn't have good performance in 
> acceleration.
>
> One of the testcases even run over 4 hours without finishing.
>
> However in CPU version, it just took 33 minutes. (under 16 processes)
>
> I want to know that is this a standard situation?
> or it is related to input file?
>
> thanks
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20140409/3ee7df33/attachment.html>