Re: [petsc-users] seg fault with superlu

thanks, using " -mat_superlu_dist_fact SamePattern_SameRowPerm" works also with options other than  "-mat_superlu_dist_colperm NATURAL". The solution worked for 32bit integers but when I compile petsc using 64bit indices I get the following error regardless of the options used for 

Re: [petsc-users] How can I get the global integer ids in a label?

Hi, Matt: Thank you for your helps. The exodus file which I used is generated by PointWise. I should try to make sure whether I can get a periodic topology using PointWise.Can you point me out which software can generate an exodus file with periodic topology? Thanks. leejearlOn Thu,

Re: [petsc-users] How can I get the global integer ids in a label?

Hi, Matt: I find that the periodic boundary is specified in the following way DMPlexCreateBoxMesh(comm, dim, user->simplex, user->cells, NULL, NULL, user->periodicity, interpolate, dm). If I use an exo file, how does the periodicity is specified? Thanks LeejearlOn Thu, 2018-12-06 at 21:42

Re: [petsc-users] How can I get the global integer ids in a label?

Hi Matt:My mesh file is "*.exo". Can it works fine?Thanks.leejearl On Thu, 2018-12-06 at 21:20 -0500, Matthew Knepley wrote: > On Thu, Dec 6, 2018 at 9:08 PM leejearl > wrote: > > Hi, Matt: > > Is the code in following page? > > > >

Re: [petsc-users] How can I get the global integer ids in a label?

Hi, Matt: Is the code in following page? /petsc-master/src/snes/examples/tutorials/ex12.c.html Thanksleejearl On Thu, 2018-12-06 at 20:26 -0500, Matthew Knepley wrote: > On Thu, Dec 6, 2018 at 8:06 PM leejearl > wrote: > > Hi Matt: > > Thank you for your helps. I want to

Re: [petsc-users] How can I get the global integer ids in a label?

Hi, Matt: Thanks for your helps. I will tell you if the code of "SNES ex12" can solve my problem.leejearl On Thu, 2018-12-06 at 20:26 -0500, Matthew Knepley wrote: > On Thu, Dec 6, 2018 at 8:06 PM leejearl > wrote: > > Hi Matt: > > Thank you for your helps. I want to implement the

Re: [petsc-users] How can I get the global integer ids in a label?

Hi Matt:Thank you for your helps. I want to implement the periodic boundary in dmplex. For a cell or face, the donor cell or face might be not distributed in the same local dm. Maybe I should ask that howcan I implement the topology of the periodic boundary in demplex.Thanks again

Re: [petsc-users] seg fault with superlu

Thanks for the suggestion I will try this option. I used already "-mat_superlu_dist_colperm NATURAL" which seems to help but I did not test it thoroughly.   Betreff: Re: [petsc-users] seg fault with superlu   On Thu, Dec 6, 2018 at 2:55 AM Matthew Knepley via petsc-users

Re: [petsc-users] seg fault with superlu

I am already using superlu_dist 6.1 but I  used to "--download-superlu_dist=1 --download-superlu_dist-commit=001d994a22684fd60cab5530cc2192f19fc58e83" when configuring petsc, is this ok?   Don't remember - but you can try with latest superlu_dist and see if the problem persists. [it has a

[petsc-users] PETSc Data structures on GPU

For parallel computations on the CPU, we allocate our own vectors and then give PETSc a point to our vectors using VecCreateMPIWithArray. For computations only on the GPU, we allocate vectors on the GPU and we’d like to do the same thing and give PETSc a pointer to our vector. 1. Does

Re: [petsc-users] PETSc binary write format

I'm attaching the file vector.dat and the file petsc saves (vec_old.dat) with this email. Thank You, Sajid Ali Applied Physics Northwestern University vector.dat Description: Binary data vec_old.dat Description: Binary data

[petsc-users] How can I get the global integer ids in a label?

Hi, Petsc developer: I have a problem for helps. There is a dmplex object, and I have distribute it use the routine DMPlexDistribute(). Then, I want to get the global integer ids in the label "Face Sets". The code is as follow, ierr = DMGetLabelIdIS(dm, "Face Sets",

Re: [petsc-users] seg fault with superlu

Don't remember - but you can try with latest superlu_dist and see if the problem persists. [it has a bunch of fixes] To try latest superlu_dist - you can use balay/update-superlu_dist_6.1.0/maint_no_merge [this branch is off maint - so to try with master - you can merge this with master] git

Re: [petsc-users] How to solve multiple linear systems in parallel, one on each process?

I was wrong with the decomposed case structure: In case of a multicore simulation, I get a decomposed case with local data (matrix+rhs+x, neighbour processors (==rank) numbers) on each process/rank. There are no halo regions (instead the original application works with data streams to