expose setting of max-steps to python interface & test#581
expose setting of max-steps to python interface & test#581speth merged 1 commit intoCantera:masterfrom
Conversation
Codecov Report
@@ Coverage Diff @@
## master #581 +/- ##
=========================================
+ Coverage 67.69% 67.7% +0.01%
=========================================
Files 362 362
Lines 38456 38459 +3
=========================================
+ Hits 26031 26039 +8
+ Misses 12425 12420 -5
Continue to review full report at Codecov.
|
|
Hey @arghdos! Thanks for this. Can you also add a method to get the max steps as well? Does CVODES expose that information? |
|
@bryanwweber Hmm, it doesn't look like they allow you to query the maximum # of steps, however we should be able to just return whatever value is stored in the integrator. |
|
@bryanwweber -- implemented |
speth
left a comment
speth
left a comment
There was a problem hiding this comment.
I had just a few minor suggestions, but this generally looks fine to me.
skyreflectedinmirrors
force-pushed
the
expose_maxsteps
branch
2 times, most recently
from
56758bb to
834cc72
Compare
skyreflectedinmirrors
force-pushed
the
expose_maxsteps
branch
from
834cc72 to
80ae4fe
Compare
|
Sorry, I didn't notice that you updated this. LGTM! |
3 participants
Changes proposed in this pull request:
I am using Cantera (as typically) to validate other ODE solvers. As such I typically want to set the network reactor tolerances to very stringent values (e.g.,
atol=1e-20,rtol=1e-15) and run ignition problems over a wide range of conditions.The issue I'm currently facing is that the default # of steps (appears to be 20k?) works for most cases, but not all. Rather than playing the game of trying to modify the tolerances to fit in the number of steps, it would be nice if we could just expose this property to the user via the python interface.
Let me know if there are any comments / concerns,
Best,
Nick