You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Anderson Acceleration script doesn't work with rpointer updates #3575
After rpointer files were updated to include year-month-date in their names the logic applied in Anderson Acceleration doesn't work. This is ultimately a change that needs to be done in the Anderson acceleration code, but I'm creating a CLM side issue because it influences MIMICS usability in more recent tags and is a blocker for using MIMICS after ctsm5.2.027. I'd like to resolve this before the CLM6 release.
@wwieder where is the Anderson code for this? In CTSM or elsewhere? Maybe it should be moved into CTSM if not, and if we want it to work and continue to work, we need to add at least a simple test for it.
Yes, sustainable use of the Anderson code is another topic. Currently it's on another repo and I need to communicate with collaborators about how to make this usable by the community.
@wwieder and I worked on getting this code off of scratch (where Will's version seemed to have become corrupted). Will now has a pull request to troubleshoot from.
Will got past this problem, which turned out to be a library-loading issue. Now that this runs, Will is confirming whether things look reasonable. I will close the specific issue as resolved.
Activity
ekluzek commentedon Oct 28, 2025
@wwieder where is the Anderson code for this? In CTSM or elsewhere? Maybe it should be moved into CTSM if not, and if we want it to work and continue to work, we need to add at least a simple test for it.
wwieder commentedon Oct 28, 2025
Yes, sustainable use of the Anderson code is another topic. Currently it's on another repo and I need to communicate with collaborators about how to make this usable by the community.
slevis-lmwg commentedon Dec 3, 2025
@wwieder and I worked on getting this code off of scratch (where Will's version seemed to have become corrupted). Will now has a pull request to troubleshoot from.
slevis-lmwg commentedon May 20, 2026
Will got past this problem, which turned out to be a library-loading issue. Now that this runs, Will is confirming whether things look reasonable. I will close the specific issue as resolved.