mathias
Forum Replies Created
-
Posts
-
Dear Shreyas,rnrn It is not working with “”blockGeometry.setNx ( some new nX)””. You can do it like this: rnrntemplate<typename T>rnvoid STLreader<T>::read(BlockGeometry3D &matrix, unsigned direction,rn unsigned voxelNumber, unsigned pad,rn double fraction, unsigned samples,rn unsigned offsetXN, unsigned offsetXP,rn unsigned offsetYN, unsigned offsetYP,rn unsigned offsetZN, unsigned offsetZP) const {rnrnrnwhere you distinguish between positive P and negative N direction. rnrnMathias
Dear Shreyas,rnrn1) You can define and an offset (cf. read function is stlReader.h) of constant size. In the next release we will have an extension for arbitrary offsets. It is not a big thing to program it in “stlRealer.h”.rn2) I am not sure that I have understood your question correctly. Actually the no. of cuboids is defined by the user.rnrnRegardsrnMathias
Dear Philipp,rnrn This depends on the contributions of OpenLB authors and users who are willing to share their work under GPL2 license. OpenLB is an open community – so everyone is invited to participate and share! Their names will be listed on the website. Interested friends of OpenLB can send an email to contact(a)openlb.net to be involved.rnrnMathiasrn rn
Dear Bolt,rnrn Depending on actual shape of the obstracle Bounce Back might just 1st order accurate. You will get better results if you refine the mesh and a better order with Bouzidi’s boundary conditions. We are currently working on Bouzidi’s boundary conditions.rnrnMathias
Dear Patrick,rnrn Please download release 1 of 0.7 if you have not done yet. Further, please send the Makefile.inc you are using to bug@openlb.net . I am using gcc 4.4.3 and openmpi and just tested this example which was working fine.rnrnMathias
Dear Bolt,rnrn The reason is the motivation behind Bounce Back: a particle moving towards the wall, coliding with it and being reflexted. So you can also include it in your step 3. It really depends on how you have implemented your LBM, e.g. if you a working with ghost cells.rnrnMathias
Dear Bolt,rnrn I would include it in the stream step.rnrnMathias
Dear both,rnrn Unfortunately I have no experience compiling OpenLB on a Windows system. Maybe there is a problem with included opose to “”boost”” files. You can try and remove the line “”src/external/boost “” in the file “”Makefile.inc””.rnrnMathias
Dear Atul,rnrn Compiling OpenLB using linux is quite easy. First you need to install a compiler like “”gnu gcc””, and also “”make””. Both is standard and probably already installed on your system. If you want to run OpenLB in parallel you also need to install “”mpi””. Then, go into one of the example folders and type “”make””. Then, OpenLB is compiling automatically. To run the example program type ./””program name””. You will also find our user guid helpful.rnrnBest regardsrnMathias
Dear Jay,rnrn Have a look at “”cylinder2d”” and “”cylinder3d””. There is an object in a tube. We have implemented Lagrangian particles recently. It is working well and we think of releasing them in one of the next releases. You can contact me directly if you want to know more -> info(at)openlb.net.rnrnMathias
Dear Jeff,rnrnat first I would try simple Bounce Back as boundary condition for all corners. I further recomment to have a look at papers of Michael Junk et al.. I remember some very helpful papers of him about that topic.rnrnMathias
The problem should be solved now. Thanks a lot for reporting!rnMathias
Dear Atul,rnrn Unfortunately, I cannot help you – I am working with Linux. But it should be possible to run the Makefile. I heart from people working with Eclipse in Windows. You can also run a virtual machine, then you have a Linux system within Windows.rnrnMaybe soneone else of the forum community can share some experience? rnrnMathias
Dear Jepson,rnrn A few years ago we were researching this problem in detail (cf. 2009, Heuveline, V. & Krause, M.J. & Latt, J.: “”Towards a Hybrid Parallelization of Lattice Boltzmann Methods””, article in Computers and Mathematics with Applications, 58, 1071-1080). Basically we found that you need to pin the OMP-threads to specific cores to reach almost the performance obtained using MPI. Unfortunately, you need some system calls to do that. Since we want OpenLB to be as generic as possible, we decided against these optimizations. rnrnIn the mentioned paper you will also find some benchmark results to which you can compare yours. However, for a sufficient large grid size I would expect a better speed-up using 48 cores.rnrnMathias rnrn
Indeed there was a bug! We have released a bug-fixed version.rnrnThanks a lot for reporting!rnMathiasrn