New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix: QNDF mass matrix tests #1094
Conversation
Split Regression test set into two sets |
@@ -653,8 +653,6 @@ function perform_step!(integrator,cache::QNDFConstantCache,repeat_step=false) | |||
end | |||
backward_diff!(cache,D,D2,k) | |||
end | |||
else |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
what happens during the else now?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The gamma value should not be updated before cnt>k
, the code was doing that previously. You can see the same results in QNDF1. (Current Implementation gives 1st order estimate only which needs to be checked) Even the norm
value is same as QNDF1
src/caches/bdf_caches.jl
Outdated
@@ -327,7 +327,7 @@ function alg_cache(alg::QNDF,u,rate_prototype,uEltypeNoUnits,uBottomEltypeNoUnit | |||
|
|||
max_order = 5 | |||
atmp = similar(u,uEltypeNoUnits) | |||
utilde = nlsolver.tmp | |||
utilde = similar(u) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
utilde = similar(u) | |
utilde = zero(u) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay
@ChrisRackauckas made the requested changes. |
Tests fail |
Rebuiliding Travis |
If you compare the times of all jobs in last master successful build here and this build, it seems like Travis has cut down it's resources recently and every job has its time increased. I think the best bet is to split up the Regression tests. |
Alright @kanav99 thanks for the help. I will update the PR with split regression tests. |
@ChrisRackauckas please have a look. |
Codecov Report
@@ Coverage Diff @@
## master #1094 +/- ##
==========================================
+ Coverage 77.33% 77.33% +<.01%
==========================================
Files 95 95
Lines 31656 31659 +3
==========================================
+ Hits 24481 24485 +4
+ Misses 7175 7174 -1
Continue to review full report at Codecov.
|
These fixes will do more than just mass matrices: check the benchmarks and see if they improved. |
Fixes All BDF and NDF methods for mass matrix tests #1077.
@ChrisRackauckas please review.