Vanishing Jacobian #245
Replies: 5 comments 5 replies
-
I am wondering if this is plausibly an issue with the version of MPI, Cuda that I am using? The modules that I load on my cluster are: |
Beta Was this translation helpful? Give feedback.
-
For anybody curious, I have solved this problem. A vanishing Jacobian error is indeed dependent on a poor choice of Intel/GCC compilers. I am not sure that providing more detail will be helpful as I imagine varies per cluster. Tony |
Beta Was this translation helpful? Give feedback.
-
Can you please explain what compilers you used for Fortran, C and C++ when it was failing. |
Beta Was this translation helpful? Give feedback.
-
Please contact me by email.
…On Fri, Mar 19, 2021 at 11:00 AM Sean Crosby ***@***.***> wrote:
@stgeke <https://github.com/stgeke> Any way we could get the ldd output
from your working install?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#245 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AEN2M772JG5BGG23ND4XFSLTEMODBANCNFSM4Y73OKCA>
.
|
Beta Was this translation helpful? Give feedback.
-
This was related to bug fixed here: |
Beta Was this translation helpful? Give feedback.
-
I am trying to run an example case, no matter which I select, be it turbPipe or ethier, I get a vanishing Jacobian error.
If someone else is able to confirm the example cases run, my guess would be it is a precision error, I was wondering if it's possible that by including CUDA I have inadvertently switched to single-precision.
Beta Was this translation helpful? Give feedback.
All reactions