Jaron T. Krogel
Jaron T. Krogel
At least one of the energy density tests (diamond cell grid for DMC) appears to be very close to passing currently: https://cdash.qmcpack.org/tests/22015901?graph=status All the statistical tests pass, as well as...
The voronoi tests on LiH (ae and pp) have most tests passing also, but show a different pattern. All the deterministic checks pass, as well as the statistical tests on...
The VMC diamond cell grid test shows intermittent passing, unlike the others: https://cdash.qmcpack.org/tests/22015955?graph=status One reason this statistical test is less stable than others is that it involves 24 statistical checks...
The above tests are for GCC14 on nitrogen. Other environments also given intermittent passing results for the diamond VMC case: sulfur GCC14 and Clang20 nitrogen2 GCC14 and Clang20 bora Intel21
In short, the following issues need investigating (in probable priority order): - [ ] Why do scalar.dat and stat.h5 give deterministically different quantities for DMC? (likely an issue broader than...
It is worth adding a deterministic test that should pass to high precision (regardless of run length): in post-processing, sum the energy density components and verify that they match the...
Does this option affect SD and MSD equally?
I wonder if this could be related to the odd behavior @anbenali is seeing for long enough MC trajectories with MSD. Is it possible that larger MSD would accumulate errors...
This is key: > The validity of the parameter also should be determined at that level as it will require a comparison to the steps per block another driver level...