<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<div dir="ltr">
<div dir="ltr"><br clear="all">
<div>
<div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature">
<div dir="ltr"><br>
</div>
</div>
</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Wed, Mar 13, 2019 at 3:49 PM Balay, Satish <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
mailers that only format html and not the text form are annoying :(<br>
<br>
<br>
On Wed, 13 Mar 2019, Zhang, Junchao wrote:<br>
<br>
> Satish, I found something strange. I configured with --with-cuda --with-precision=single, then with -log_view, I saw<br>
> Compiled with single precision PetscScalar and PetscReal<br>
> Compiled with full precision matrices (default) <br>
<br>
This message is misleading [and should be fixed]. This feature was removed.<br>
<br>
> <br>
> I found that confused mumps. When I added --with-precision=double, I got consistent MatScalar and PetscScalar precision, and the errors I reported<br>
> disappeared. Do you know why?<br>
<br>
how is mumps confused? Perhaps this example doesn't work with single precision? [or mumps single precision doesn't work for this problem?]<br>
</blockquote>
<div> </div>
<div> OK, I thought mumps was inputed with a double precision matrix and single precision vectors. I was wrong. I tested on my laptop --with-cuda --with-precision=single, these errors did show up. In petsc dashboard, we do not have "single + mumps" combination.
I will create a PR to mark these tests require double.</div>
<div><br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Lots of examples fail with single precision..<br>
<br>
Satish<br>
<br>
> <br>
> --Junchao Zhang<br>
> <br>
> <br>
> On Wed, Mar 13, 2019 at 11:15 AM Junchao Zhang <<a href="mailto:jczhang@mcs.anl.gov" target="_blank">jczhang@mcs.anl.gov</a>> wrote:<br>
> I met some errors with cuda + mumps. It was tested with<br>
> make -f gmakefile test search='snes_tutorials-ex69_q2p1fetidp_deluxe snes_tutorials-ex62_fetidp_2d_quad<br>
> snes_tutorials-ex69_q2p1fetidp_deluxe_adaptive ksp_ksp_tutorials-ex52f_mumps'<br>
> <br>
> I can reproduce it with petsc master. The first line of petsc nightly<br>
> (<a href="http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2019/03/12/master.html" rel="noreferrer" target="_blank">http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2019/03/12/master.html</a>) shows another error. But I guess they have the same root:
PETSc<br>
> gives random wrong results in some cases. For example, I ran ksp_ksp_tutorials-ex52f_mumps twice and saw<br>
> <br>
> $ mpirun -n 3 ./ex52f<br>
> Mumps row pivot threshhold = 1.00E-06<br>
> Mumps determinant=( 9.01E-01 0.00E+00)*2^ 99<br>
> Norm of error 1.5554E-06 iterations 1<br>
> <br>
> $ mpirun -n 3 ./ex52f<br>
> Mumps row pivot threshhold = 1.00E-06<br>
> Mumps determinant=( 9.01E-01 0.00E+00)*2^ 99<br>
> Norm of error 1.6356E-06 iterations 1 <br>
> <br>
> <br>
> The correct output has "Norm of error < 1.e-12,iterations 1". Currently, I do know the reason.<br>
> <br>
> --Junchao Zhang<br>
> <br>
> <br>
> <br>
</blockquote>
</div>
</div>
</body>
</html>