Re: bug in linearBicgstabSolver.py
no worries, thanks On Wed, Jun 22, 2016 at 10:58 AM, Daniel Wheeler wrote: > Jamie, the new code is on the develop branch now > > > https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_usnistgov_fipy_blob_develop_fipy_solvers_scipy_linearBicgstabSolver.py&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=FoPBLFKX5Z7TW3Sk-72AFrPn3EJESyjcGiyOfgsy7_s&s=Biiu3V9QVjqpBgmvVqdcLIoTV88lYbiG1-anl-4mUSI&e= > > It seemed to work for a small test case. Unfortunately, that solver > isn't part of the test suite. > > On Wed, Jun 22, 2016 at 10:42 AM, James Pringle wrote: > > thanks > > Jamie > > -- > Daniel Wheeler > ___ > fipy mailing list > fipy@nist.gov > > https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ctcms.nist.gov_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=FoPBLFKX5Z7TW3Sk-72AFrPn3EJESyjcGiyOfgsy7_s&s=J6QRk5zhR5NwcHCD_cwJ_TtPPK1BnT-89TSdrw28XPo&e= > [ NIST internal ONLY: > https://urldefense.proofpoint.com/v2/url?u=https-3A__email.nist.gov_mailman_listinfo_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=FoPBLFKX5Z7TW3Sk-72AFrPn3EJESyjcGiyOfgsy7_s&s=7RgqyeMB4ocaYTTrqqckSEauTlHtvJOG8aO9T9DBBHI&e= > ] > ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
Re: bug in linearBicgstabSolver.py
Jamie, the new code is on the develop branch now https://github.com/usnistgov/fipy/blob/develop/fipy/solvers/scipy/linearBicgstabSolver.py It seemed to work for a small test case. Unfortunately, that solver isn't part of the test suite. On Wed, Jun 22, 2016 at 10:42 AM, James Pringle wrote: > thanks > Jamie -- Daniel Wheeler ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
Re: bug in linearBicgstabSolver.py
thanks Jamie On Wed, Jun 22, 2016 at 10:04 AM, Guyer, Jonathan E. Dr. (Fed) < jonathan.gu...@nist.gov> wrote: > That's terrible. Not to name any names, but Daniel did that about five > years ago. > > I've submitted a fix at > https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_usnistgov_fipy_pull_497&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=rqSLJ5XWBtgVE-7mbJie2VpGSqiUU-ARkbMDtlhb-Qk&s=ekHWBVJYr8pw7evCYx2IbG68LgM2t0bzf9V4FO08Psw&e= > > > On Jun 21, 2016, at 9:22 PM, James Pringle wrote: > > > > The document string for fipy.solvers.scipy.linearBicgstabSolver.py > claims it takes the arguments "tolerance," "iterations" and "precon", but > it does not. > > > > To reproduce this bug, make sure you are using the scipy solvers, and > compare the results for, e.g. > eq.solve(var=psi,solver=fipy.LinearPCGSolver(tolerance=1.0e-3)) to > eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)). The > later throws the error: > > > > In [43]: > eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)) > > > --- > > TypeError Traceback (most recent call > last) > > in () > > > 1 > eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)) > > > > TypeError: __init__() got an unexpected keyword argument 'tolerance' > > > > > > In fipy version 3.1; the error appears to be in the latest git version > of the code as well; it appears the problem is in the > fipy/fipy/solvers/scipy/linearBicgstabSolver.py, where the arguments are > simply not based to bicgstab. > > > > This is too bad, for LinearBicgstabSolver gives by far the best results > on my problem, and the underlying scipy.sparse.linalg.bicgstab does take > the argument "tol". > > > > I think I could fix the problem staring at the code, but the > class/inheritance that is going on is above what I really understand, and I > am afraid I would bolix things further. > > > > Thanks, > > Jamie > > ___ > > fipy mailing list > > fipy@nist.gov > > > https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ctcms.nist.gov_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=rqSLJ5XWBtgVE-7mbJie2VpGSqiUU-ARkbMDtlhb-Qk&s=53Zt7rb_fGwvi-ZERkInTL4oEOiILcAkz4LrCqREm_o&e= > > [ NIST internal ONLY: > https://urldefense.proofpoint.com/v2/url?u=https-3A__email.nist.gov_mailman_listinfo_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=rqSLJ5XWBtgVE-7mbJie2VpGSqiUU-ARkbMDtlhb-Qk&s=xYOeFF5MIcAEsgelMexvlh_smIT-Bb36TV2ypQPDllg&e= > ] > > > ___ > fipy mailing list > fipy@nist.gov > > https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ctcms.nist.gov_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=rqSLJ5XWBtgVE-7mbJie2VpGSqiUU-ARkbMDtlhb-Qk&s=53Zt7rb_fGwvi-ZERkInTL4oEOiILcAkz4LrCqREm_o&e= > [ NIST internal ONLY: > https://urldefense.proofpoint.com/v2/url?u=https-3A__email.nist.gov_mailman_listinfo_fipy&d=CwICAg&c=c6MrceVCY5m5A_KAUkrdoA&r=7HJI3EH6RqKWf16fbYIYKw&m=rqSLJ5XWBtgVE-7mbJie2VpGSqiUU-ARkbMDtlhb-Qk&s=xYOeFF5MIcAEsgelMexvlh_smIT-Bb36TV2ypQPDllg&e= > ] > ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
Re: bug in linearBicgstabSolver.py
That's terrible. Not to name any names, but Daniel did that about five years ago. I've submitted a fix at https://github.com/usnistgov/fipy/pull/497 > On Jun 21, 2016, at 9:22 PM, James Pringle wrote: > > The document string for fipy.solvers.scipy.linearBicgstabSolver.py claims it > takes the arguments "tolerance," "iterations" and "precon", but it does not. > > To reproduce this bug, make sure you are using the scipy solvers, and compare > the results for, e.g. > eq.solve(var=psi,solver=fipy.LinearPCGSolver(tolerance=1.0e-3)) to > eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)). The > later throws the error: > > In [43]: eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)) > --- > TypeError Traceback (most recent call last) > in () > > 1 eq.solve(var=psi,solver=fipy.LinearBicgstabSolver(tolerance=1.0e-3)) > > TypeError: __init__() got an unexpected keyword argument 'tolerance' > > > In fipy version 3.1; the error appears to be in the latest git version of the > code as well; it appears the problem is in the > fipy/fipy/solvers/scipy/linearBicgstabSolver.py, where the arguments are > simply not based to bicgstab. > > This is too bad, for LinearBicgstabSolver gives by far the best results on my > problem, and the underlying scipy.sparse.linalg.bicgstab does take the > argument "tol". > > I think I could fix the problem staring at the code, but the > class/inheritance that is going on is above what I really understand, and I > am afraid I would bolix things further. > > Thanks, > Jamie > ___ > fipy mailing list > fipy@nist.gov > http://www.ctcms.nist.gov/fipy > [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ] ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
