More application testing energy density estimator#6011
Conversation
|
These failing checks are all failures on the CI side. I think #6016 may address one but something with the coverage is also broken. |
prckent
left a comment
prckent
left a comment
There was a problem hiding this comment.
LGTM
Good to have short batched tests of these where we previously had none.
ye-luo
left a comment
ye-luo
left a comment
There was a problem hiding this comment.
Only a few clarification questions on check_stats.py.
| sc_vals = scalars.data[comparisons[k]] | ||
| if scalars.file_type=='dmc': | ||
| if len(sc_vals)%len(ed_vals)==0 and len(sc_vals)>len(ed_vals): | ||
| print("len(sc_vals)%len(ed_vals)", len(sc_vals)%len(ed_vals), len(sc_vals), " >= ", len(ed_vals)) |
There was a problem hiding this comment.
Is this line for debugging or intended to be visible to users? See quite casual printout.
| sc_vals = sc_vals.sum(1) | ||
| sc_weights = sc_weights.reshape(len(sc_weights)//steps,steps) | ||
| sc_weights = sc_weights.sum(1) | ||
| sc_vals = sc_vals / sc_weights |
There was a problem hiding this comment.
It seems to be an important bug fix getting proper weighted average now. Did I understand it correct? I didn't find anything mentioned in the PR description. So wondering.
| # This test is actually relative error assuming the | ||
| # the scv is the true value. since conventionally the | ||
| # divisor is the true value. | ||
| print("checking estimator", comparisons[k], "versus scalar agreement") |
There was a problem hiding this comment.
Is this just for debugging?
| if scalars.file_type=='dmc': | ||
| ftol = 1e-7 | ||
| else: | ||
| ftol = 1e-7 |
There was a problem hiding this comment.
Same value for if and else?
3 participants
Proposed changes
Additional Testing of Energy Density Estimator
Added:
Homogenous Electron Gas tests for batched vmc and dmc.
The energy density should be approximately homogenous.
The energy should be the same whether the coulomb AA and other hamiltonian operators should produce the same results whether batched per particle eval is taken or not. It is taken when the energy density estimator is on
What type(s) of changes does this code introduce?
Does this introduce a breaking change?
What systems has this change been tested on?
osx dev laptop
Checklist