[gmx-users] 2020.1: comm broken with -update gpu -bonded gpu ?

[Paul bauer]

Hello,

sorry for the confusion, but I'm currently moving our infrastructure 
over to Gitlab and have set Redmine to read only because of this.

I'll update gmx-users and gmx-developers once the process is done, and 
then you can open the issue on Gitlab using the tracking system there.

Cheers

Paul

On 23/03/2020 19:12, Jochen Hub wrote:
> Hi Magnus and Justin,
>
> Am 23.03.20 um 13:52 schrieb Magnus Lundborg:
> > Hi Jochen,
> >
> > Have you tested if this happens with -update cpu? Perhaps it's the 
> bonded on GPU that's the problem, unrelated to updating.
>
> The problem only appears with -update gpu, not with -update cpu.
>
>  I didn't see a
> > Redmine issue yet. Could you make one?
>
> I tried, but I seem to lack permissions: On
>
> https://redmine.gromacs.org/projects/gromacs/issues/new
>
> I get "You are not authorized to access this page." Did I miss 
> something here?
>
> Thanks,
> Jochen
>
>
> Am 23.03.20 um 13:52 schrieb Magnus Lundborg:
>> Hi Jochen,
>>
>> Have you tested if this happens with -update cpu? Perhaps it's the 
>> bonded on GPU that's the problem, unrelated to updating. I didn't see 
>> a Redmine issue yet. Could you make one?
>>
>> Cheers,
>> Magnus
>>
>> On 2020-03-19 18:01, Jochen Hub wrote:
>>> Hi developers,
>>>
>>> I am running a simple DPPC membrane (Berger force field, PME, 1nm 
>>> cutoff, 4fs time step, all standard) with 2020.1 and
>>>
>>> comm-mode                = Linear
>>> nstcomm                  = 100
>>> comm-grps                = DPPC water  ; OR System
>>>
>>> but the membrane rapidly drifts along the z direction: approx. once 
>>> across the box per 100ps, and accelerating over time.
>>>
>>> This happens only with
>>>
>>> mdrun -update gpu -bonded gpu
>>>
>>> but not with
>>>
>>> mdrun -update gpu -bonded cpu   (no spelling mistake)
>>>
>>> (with a GTX 1070Ti).
>>>
>>> Also no problems with 2018.6 or 2019.6.
>>>
>>> Seems like the center of mass motion removal is broken when doing 
>>> both *bonded and updating* on the GPU. Is this issue known?
>>>
>>> Cheers,
>>> Jochen
>>>
>>
>

-- 
Paul Bauer, PhD
GROMACS Development Manager
KTH Stockholm, SciLifeLab
0046737308594