Hi Lisandro, I haven't checked this out yet, but I'm going to guess that this is probably my fault. The CSRPERM matrix types were added by me some time ago and I haven't maintained them in a while. They are currently broken because I didn't keep up with some of the changes to how PETSc matrices work. I will hopefully find the time to fix this soon, as I have some much improved versions of the CSRPERM stuff lying around, but I just haven't gotten the time to clean things up and push to petsc-dev.
--Richard Lisandro Dalcin wrote: > Can someone confirm if this fails in 1 proc? I am not sure what's > going on, but there should be something wrong with the internal > convert machinery, > > o = Mat().create() > o.setSizes(0) > o.setType('mpicsrperm') -- Richard Tran Mills, Ph.D. Computational Scientist E-mail: rmills at ornl.gov Computational Earth Sciences Group Phone: (865) 241-3198 Oak Ridge National Laboratory Fax: (865) 574-0405