Re: [petsc-users] MatConvert changes distribution of local rows

Matrix types is from MATMPIDENSE to MATSCALAPACK, but I think it happens also for other matrix types IIRC.     Gesendet: Freitag, 13. Januar 2023 um 16:58 Uhr Von: "Pierre Jolivet" An: "Marius Buerkle" Cc: petsc-users@mcs.anl.gov Betreff: Re: [petsc-users] MatConvert changes distribution of loc

Re: [petsc-users] MatConvert changes distribution of local rows

> On 13 Jan 2023, at 9:18 AM, Marius Buerkle wrote: > > Matrix types is from MATMPIDENSE to MATSCALAPACK, OK, that’s not possible, because PETSc and ScaLAPACK use different distributions for dense matrices. > but I think it happens also for other matrix types IIRC. Which one? Thanks, Pierr

Re: [petsc-users] MatConvert changes distribution of local rows

Hi,   Ok thanks for clarifying.    It seems to work for the other mattypes at least with the lastest 3.18 version. I think i had the problem with come older version, but i can't remember the version number.   best and thanks, marius     Gesendet: Freitag, 13. Januar 2023 um 17:25 Uhr Von: 

[petsc-users] Retreiving a PetscObject

Hi, Is it possible with PETSc's API to query an objet by it's name? Is there a "global" database with all PETScObject created? Thanks, Eric -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

Re: [petsc-users] Retreiving a PetscObject

On Fri, Jan 13, 2023 at 4:24 AM Eric Chamberland < eric.chamberl...@giref.ulaval.ca> wrote: > Hi, > > Is it possible with PETSc's API to query an objet by it's name? > > Is there a "global" database with all PETScObject created? > No, we do not store that information. What would the use case be?

Re: [petsc-users] Retreiving a PetscObject

ok, here is what I want to do: I am writing a brand new PC which needs some other PETSc objects like a Vec (for unit partition). We have a text based user file format in which the user can create arbitrary Petsc's Vec or Mat. Then, I would like the user to give the prefix of these Vec or Ma

Re: [petsc-users] Retreiving a PetscObject

On Fri, Jan 13, 2023 at 5:58 AM Eric Chamberland < eric.chamberl...@giref.ulaval.ca> wrote: > ok, here is what I want to do: > > I am writing a brand new PC which needs some other PETSc objects like a > Vec (for unit partition). > > We have a text based user file format in which the user can creat

[petsc-users] Nonconforming object sizes using TAO (TAOBNTR)

Hi all, In a variational approach of brittle fracture setting, I try to solve a bound constraint minimization problem using TAO. I checkout on the main branch of Petsc. Minimization with respect to the bounded variable (damage) is achieved through the Bounded Newton Trust Region (TAOBNTR). All

Re: [petsc-users] Nonconforming object sizes using TAO (TAOBNTR)

On Fri, Jan 13, 2023 at 11:22 AM Alexis Marboeuf wrote: > Hi all, > > In a variational approach of brittle fracture setting, I try to solve a > bound constraint minimization problem using TAO. I checkout on the main > branch of Petsc. Minimization with respect to the bounded variable (damage) > i

Re: [petsc-users] Nonconforming object sizes using TAO (TAOBNTR)

Hi Matt, Here is the output from Petsc when I view the TAO solver: Tao Object: (Damage_) 4 MPI processes type: bntr Tao Object: (Damage_tao_bnk_cg_) 4 MPI processes type: bncg CG Type: ssml_bfgs Skipped Stepdirection Updates: 0 Scaled gradient steps: 0

Re: [petsc-users] Nonconforming object sizes using TAO (TAOBNTR)

On Fri, Jan 13, 2023 at 3:21 PM Alexis Marboeuf wrote: > Hi Matt, > > Here is the output from Petsc when I view the TAO solver: > Okay, there is no sophisticated caching going on. So, first I would get it to fail on1 process. It should if it just depends on the convergence (I hope). Then send th

Re: [petsc-users] Nonconforming object sizes using TAO (TAOBNTR)

Hi Matt, Indeed, it fails on 1 process with the same error. The source code is available here: https://github.com/bourdin/mef90 (branch marboeuf/vdef-tao-test) [https://opengraph.githubassets.com/8f51eb183957c4e2f2dd59e2733f43a7bc667a50d4aaad934ebb3ac8f25a17ab/bourdin/mef90]