Re: [ESPResSo-devel] incorrect velocity scaling for slip lbboundary

[Kai Szuttor]

Hi Ivan,

This issue is on our agenda. I already had a look but could not find the source 
of the wrong velocity.
We hope to be able to resolve it soon.

Best,

Kai

> On 25. Jan 2018, at 14:43, Ivan Cimrak <address@hidden> wrote:
> 
> Hi guys,
> 
> 
> In December I posted here comment on incorrect scaling of velocity at 
> lbboundary. Kai checked this on GPU, it was correct, however for CPU it is 
> still wrong. Please see my mail below. I am adding a sample script.
> 
> 
> Thanks,
> 
> Ivan
> 
> 
> On 08/12/2017 11:08, Kai Szuttor wrote:
>> Hi Ivan,
>> 
>> I just checked your test with GPU LB and there the velocity comes out
>> correct, so it seems to be a problem of CPU LB.
>> 
>> I'll have a look.
>> 
>> Cheers,
>> 
>> Kai
>> On Fri, Dec 08, 2017 at 08:56:03AM +0000, Ivan Cimrak wrote:
>>> Hey guys,
>>> 
>>> 
>>> Is it possible to have incorrect velocity scaling for slip boundary
>>> conditions? I took an example script (attached) from section "10.10. Setting
>>> up boundary conditions" from ESPResSo online documentation:
>>> 
>>> http://espressomd.org/html/doc/lb.html
>>> 
>>> I added the printing of the velocity near the right slip wall and after
>>> reaching steady state it shows almost 100 times larger velocity as it is
>>> supposed to be (0.941 or so instead of 0.01). This behaviour is linked with
>>> lb timestep which was 0.01. Increasing to timestep=0.1 gives only 10times
>>> higher velocity.
>>> 
>>> 
>>> Could you comment on this?
>>> 
>>> 
>>> Many thanks,
>>> 
>>> Ivan
>>> 
>> 
> 
> <test_from_documentation.py>

signature.asc
Description: Message signed with OpenPGP