Support symmetric backends across the core algorithms#1

Merged

mtfishman merged 14 commits into

mf/itensorbase-backendfrom

Jul 7, 2026

Commits on Jul 7, 2026

Use graded-capable matrix functions for gauging, entanglement, and the compat eigen

mtfishman
and
claude
committed
Support symmetric backends in inner, rdm, expect, and sampling
mtfishman
committed
Support symmetric backends in loop corrections and re-include the example tests
mtfishman
committed
Rewrite symmetric_gauge in the BP perspective with rename-based absorption

mtfishman
and
claude
committed
Pin the ITensorBase and TensorAlgebra branches while the symmetry support is in review

mtfishman
and
claude
committed
Tidy the ITensors compat: sim, onehot, and the identity map

mtfishman
and
claude
committed
Make the order-2 compat identities device-following
mtfishman
committed
Use the compatibility layer for op in the Heisenberg-picture example
mtfishman
committed
Carry the dual arrows in the device-following identity contractions
mtfishman
committed
Match gate site indices by name so apply_gates works on graded backends
mtfishman
committed
Build order-2 identities via id/one and drop the order-2 delta method
mtfishman
committed
Use name-based namesetdiff for index set differences in qr, auxinds, and full_update
mtfishman
committed
Inline balanced SVD in the apply routines and simplify compat truncation
mtfishman
committed
Remove [sources] pins now that the upstreams have registered
mtfishman
committed