Test references are outdated with respect to current gradient cutoff default parameters

[ohearnk]

Many test reference outputs under tests/saved/*.out are either inconsistent across the test categories or not aligned with current cutoff defaults (or both).

Specfically, the following are the current cutoff defaults:

• Exchange correlation (XCCUTOFF): 1e-7
• SCF density matrix entry RMS change (DENSERMS): 1e-6
• SCF basis sets (BASISCUTOFF): 1e-6
• SCF one- and two-electron integrals (CUTOFF): 1e-7
• Gradients (GRADCUTOFF): 1e-7

These are also printed in the reference test outputs under the following headers, respectively:

• PRUNING CUTOFF (XC Quadrature Formation section)
• DENSITY MATRIX MAXIMUM RMS FOR CONVERGENCE (JOB CARD section)
• BASIS FUNCTION (JOB CARD section)
• TWO-e INTEGRAL (JOB CARD section)
• GRADIENT CUTOFF (JOB CARD section)

Tests references need to be updated and validation parameters reviewed in light of current code cutoff defaults (see calls to dacdif in the runtest script for the latter point).

[ohearnk]

After reviewing this further, the main issue is that the gradient cutoffs in the test references are actually tighter (1e-9) than the current code default (1e-7) for the majority of the 83 gradient tests.

Gradient test outputs need to be updated with the newer looser default cutoff, particularly as the gradient test inputs do not override the default value of GRADCUTOFF of 1e-7.

As to the remaining cutoffs, a quick summary for the 196 tests is below for manually specified values in test inputs:

• XCCUTOFF (1e-12), BASISCUTOFF (1e-12):
  • test/esp_charge_acetate_anion.in
  • test/esp_charge_acetate_radical.in
  • test/esp_charge_acetone_b3lyp_def2svp.in
  • test/esp_grid_acetone_b3lyp_def2svp.in
• DENSERMS (1e-6, which is the current default):
  • all 196 tests
• CUTOFF (1e-10):
  • test/ene_psb3_libxc_gga_631g.in
  • test/ene_psb3_libxc_hgga_631g.in
  • test/ene_psb3_libxc_lda_631g.in
• CUTOFF (1e-12):
  • test/esp_charge_acetate_anion.in
  • test/esp_charge_acetate_radical.in
  • test/esp_charge_acetone_b3lyp_def2svp.in
  • test/esp_grid_acetone_b3lyp_def2svp.in
• CUTOFF (1e-9):
  • all remaining 188 tests except test/grad_c10h9o2s_b3lyp_def2svp.in

[ohearnk]

For reference, #437 was the genesis for bringing up this issue.