-
Notifications
You must be signed in to change notification settings - Fork 440
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
Libint2 one electron integrals #2388
Conversation
This PR isn't marked "ready for review." Is there anything this needs from other devs? |
I may need some help building a new Windows L2 library with the extra OEI hessians; the VM that I had set up is no longer available after the IT folks kindly sent a security patch that bricked my laptop. I should be able to get Parallels running again if needed, but would appreciate help from anyone with a Windows machine. The failure in ADCC is something I see locally too; I get |
Some of the |
I have a Windows box that I can try to build L2 on. |
None of my ADCC tests are passing; all have the same problem. It's an old psi4dev Conda environment, so I should probably remake that first. |
adcc/libtensorlight were among those that needed a rebuild late Nov when I moved the mkl pinning from |
Yes, that's what I meant. On Azure, only one |
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.
lgtm, though I'd treat myself as a grey checkmark on this one. delightful delta LOC.
I'll run a tests w/threading on this locally and report back.
d892a48
to
f35c00a
Compare
The latest CMake fix worked, so Linux is in good shape now. Before we can merge this, we'll need to make sure we have conda L2 set up correctly. We don't need the high angular momentum multipoles - I have a different strategy in mind for those. We will need the second derivatives of the one electron ints though. The lack of those hessian ints is what's causing the remaining failures on Windows. |
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.
Found some spots where maybe #include "psi4/libmints/osrecur.h"
is not needed anymore 😄
Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu>
Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu>
Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu>
Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu>
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.
I feel a lot better about this, knowing that I can bug either Andy or Max on this PR if we have problems later. Please tend to the two changes I requested in this latest round, and again, reminder to coordinate with Lori on where to get L2 from, for test purposes.
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.
Thanks for addressing the comments! Once L2 conda deployment + tests are working, good to go! 🚀
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.
Thank you for all of the hard work! LGTM
|
||
#define MAX(a, b) ((a) > (b) ? (a) : (b)) |
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.
#define MAX(a, b) ((a) > (b) ? (a) : (b)) |
|
||
#define MAX(a, b) ((a) > (b) ? (a) : (b)) |
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.
#define MAX(a, b) ((a) > (b) ? (a) : (b)) |
|
||
#define MAX(a, b) ((a) > (b) ? (a) : (b)) | ||
|
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.
#define MAX(a, b) ((a) > (b) ? (a) : (b)) |
* Adds .gitattributes file to normalize line endings. * Starting on libint2 overlaps * Add devcontainer stuff * Working on converting one-electron integrals to * Full ctests passing at this point * Fixes for Pytests. All tests pass now. * Remove buggy Pseudospectral derivs, GShell functions * Make compute_pair private again * Fix and parallelize PCM integrals * Get correct l2 library when testing - to be reverted later * Remove Git setup files * Get L2 from the right repo * Remove buffer call from one body ints * Fix syntax in matrix header * Correct for antisymmetric operators in onebody compute method * Make sure CMake applies same standard to all L2 files * Add Lori suggestions and docs, remove more physconst includes * Use make_unique in one electron ints * Use libint2 for relativistic potential ints * Fix PCM with GCC * Update azure-pipelines-windows.yml * Update psi4/src/psi4/libmints/dipole.cc Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu> * Update psi4/src/psi4/libmints/dipole.cc Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu> * Update doc/sphinxman/source/prog_integrals.rst Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu> * Update doc/sphinxman/source/prog_integrals.rst * Update doc/sphinxman/source/prog_integrals.rst Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu> * Update doc/sphinxman/source/prog_integrals.rst * Use pairlist for gradient and hessian integrals * Add more docs * Fix test_mints pytest * Apply changes requested in reviews * Remove test of ADC with symmetry * Small review comments * Add Lori's suggested test fix * Update meta.yaml Co-authored-by: Justin Turney <justin.turney@gmail.com> Co-authored-by: Lori A. Burns <lori.burns@gmail.com> Co-authored-by: Jonathon Misiewicz <jonathon.paul.misiewicz@emory.edu>
Thank you so much for this guys! This has been a real sticking point for me for a lot of applications of Psi4. |
Description
Adds libint2 one electron integrals. A few integrals still use the hand-written code for now, but they will be addressed in subsequent PRs, as linked below.
Todos
New Features for Release Notes
Checklist
Status