All issues

Search results

TD gradient OOMs with 34 GB free: _jk_energies_by_dm_factors commits more than get_avail_mem reports

Status: Open.

#881 In pyscf/gpu4pyscf;

· michalkrompiec opened on Aug 30, 2026

[Bug]: ccsd_incore returns incorrect, non-deterministic energies on Blackwell GB10 (sm_121) for 6-31G*

Status: Open.

#880 In pyscf/gpu4pyscf;

· chixson opened on Aug 29, 2026

Possibly expensive QM/MM energy_nuc is computed for every SCF cycle

Status: Open.

#872 In pyscf/gpu4pyscf;

· zc62 opened on Aug 25, 2026

Non-equilibrium TDDFT solvation hardcodes eps = 1.78, silently discarding the user's solvent

Status: Open.

#867 In pyscf/gpu4pyscf;

· michalkrompiec opened on Aug 21, 2026

Make pyscf-dispersion optional dependency as in PySCF

Status: Open.

#866 In pyscf/gpu4pyscf;

· awvwgk opened on Aug 21, 2026

CuTensor bug with cutensor-cu12==2.7.0 when running UKS density fitting hessian

Status: Open.

#850 In pyscf/gpu4pyscf;

· henryw7 opened on Aug 11, 2026

[BUG REPORT] nao mismatch causes RSJK PBC DFT failures with contracted basis sets

Status: Open.

#846 In pyscf/gpu4pyscf;

· lhan67 opened on Aug 3, 2026

PCM analytical IR dipole disagrees with finite-difference PCM dipoles

feature request

Status: Open.

#844 In pyscf/gpu4pyscf;

· corinwagen opened on Jul 31, 2026

Hessian raises assert *.flags.c_contiguous after upgrading 1.7.6 → 1.8.0

Status: Open.

#836 In pyscf/gpu4pyscf;

· jevandezande opened on Jul 25, 2026

DF.reset() keeps a stale auxmol after in-place geometry change — silently wrong SCF energies (diverges from PySCF CPU behavior)

Status: Open.

#827 In pyscf/gpu4pyscf;

· Isitea opened on Jul 22, 2026

Excessive temporary allocation in TDDFT due to cp.einsum contraction path

Status: Open.

#821 In pyscf/gpu4pyscf;

· 2azyking opened on Jul 14, 2026

mf.omega override on range-separated functionals gives wrong energies on GPU (works on CPU)

Status: Open.

#815 In pyscf/gpu4pyscf;

· jevandezande opened on Jul 8, 2026