Changelog

Added

• A multi-IR staged pass manager (#16418)
• Allow 32-bit registers in classical.expr system (#16449)
• Add C API for inspecting control flow instructions (#16179)
• Add register query functions to the C API (#16177)
• Var-based measurement experiments (#16273)
• Add C API for inspecting classical expressions (#16178)
• Use GateCount over Depth in Clifford+T pipeline (#16218)
• Add QuantumCircuit method to compute estimated fidelity (#16292)
• Add C API for two qubit unitary peephole optimization pass (#16088)
• Extend litinski transformation for all Pauli rotations (#15974)
• Add additional Z and X rotation gate handling to CommutativeCancellation (#16191)
• Add PBC transpilation pipeline (#16270)
• Add Custom Gates/Instructions in Rust (#15684)
• Add QubitSparsePauliList.commutes (#16194)
• extend ConvertToPauliProductRotations to PPRs and PPMs (#16306)
• Expose qk_{circuit, dag}_global_phase in the C API (#16278)
• Add generate_clifford_t_pass_manager (#15918)
• Make uuid public on ParameterVector (#16216)
• Add DAG-based variant of BasisTranslator transpiler pass in C API (#15783)
• Add new multithreaded TwoQubitPeepholeOptimization pass (#13419)
• Improve single-qubit gate count in TwoQubitControlledUDecomposer (#16123)
• Track PauliLindbladMap generators during parity_sample (#16072)
• Add auto-generated ctypes bindings to C API (#16067)
• make plot_bloch_vector accept tuple input (#16103)
• Improve commutation checking of Pauli product rotations and measurements (#15815)
• Add option to set the seed_transpiler from environment (#16001)
• Enable try_matrix & friends for PPR and PPM (#16011)
• Add control flow support to ConvertToPauliRotations (#15941)
• Convert Sabre's lookahead to layer-based tracking (#14911)
• Support commutation check between Pauli-based gates and standard gates (After reverting #15488) (#15925)
• Extend RemoveIdentityEquivalent and CommutativeOptimization with PauliProductRotations (#15810)
• Support commutation check between Pauli-based gates and standard gates (#15488)
• Support phase gates in LitinskiTransformation (#15773)

Fixed

• Add duration for reset to GenericBackendV2 (#16442)
• Fixes #16166: Preserve global_phase in dagdependency_to_circuit (#16323)
• Have SynthesizeRZRotations clear rsgridsynth global caches (#16392)
• Correctly combining control modifiers in OptimizeAnnotated (#16428)
• Fix stack-overflow segfault in OpenQASM 2 expression parsers (#16421)
• Fix MultiplierGate truncated result decomposition (#16394)
• Fix CommutationChecker for controlled rotation gates (#16337)
• Fix BitArray bitcount padding handling (#16309)
• Fix determinism in ConsolidateBlocks (#16237)
• Fix redundant unit numerator in pi_check pi fractions (#16175)
• Remove QPY warning on deserialising incomplete ParameterVectors (#16222)
• Fix composing virtual permutation layouts for ElidePermutations and StarPreRouting (#13945)
• QPY: Pauli Product Measurement/Rotation parameter fix (#16294)
• Reject loopback edges in CouplingMap (#16310)
• fix: Ensure there is no subtraction panic via underflow (#16249)
• Fix panic on 1.0 error in VF2Layout (#16285)
• Fix ConstrainedReschedule barrier handling (#16151)
• Initialize ConstrainedReschedule.target unconditionally (#16246)
• Sorts gate array to have deterministic gate selection (#16139)
• Fix UnrollForLoops global phase accumulation (#16201)
• Fix Statevector.evolve modifying input state in-place (Fixes 15750) (#15905)
• Fix Commuting2qGateRouter global phase handling (#16215)
• Fix InverseCancellation for symmetric pairs (#16153)
• Adding LinearFunction.inverse (#16203)
• Fix invalid simplification of XXPlusYY gates in CommutativeOptimization (#16163)
• Add safe shape checking to synth_cz_depth_line_mr (#16154)
• fix(library): clifford_6_4 template missing global_phase causes silent rejection (#15944)
• make plot_bloch_vector accept tuple input (#16103)
• Fix QPY loading delay integers durations incorrectly (#16076)
• Fix forced block consolidation for incomplete basis sets (#16075)
• Add option to set the seed_transpiler from environment (#16001)
• Fix panic in RemoveIdentityEquivalent with ParameterVector global phase (#16054)
• Fixed commutation checking between two Pauli product measurements (backport #16023) (#16040)
• Fixed commutation checking between two Pauli product measurements (#16023)
• Avoid triggering register/layout warning in remove_final_measurements (#16026) (#16037)
• Avoid triggering register/layout warning in remove_final_measurements (#16026)
• Fix C transpiler level 3 optimisation loop (#15967) (#16008)
• Fix C transpiler level 3 optimisation loop (#15967)
• Producing correct QFT circuits over linear connectivity with 32 or more qubits (#16004) (#16007)
• Producing correct QFT circuits over linear connectivity with 32 or more qubits (#16004)
• Fix TemplateOptimization dropping the global_phase on substitution (#15943)
• Reimplementing QSD using faer instead of nalgebra. (#15874)
• Rewrite QPY ParameterExpression handling in pure Polish form (#15934) (#15961)
• Rewrite QPY ParameterExpression handling in pure Polish form (#15934)
• Support commutation check between Pauli-based gates and standard gates (After reverting #15488) (#15925)
• Correctly enforce matrix size check in CommutationChecker (#15933) (#15936)
• Correctly enforce matrix size check in CommutationChecker (#15933)
• Support commutation check between Pauli-based gates and standard gates (#15488)
• QPY Rust and Python compatibility fixes (#15847) (#15887)
• QPY Rust and Python compatibility fixes (#15847)
• Remove unwanted remapping (#15872) (#15885)
• Remove unwanted remapping (#15872)
• Fix: BasisTranslator processing of nested ControlFlowOp (#15875) (#15884)
• Fix: BasisTranslator processing of nested ControlFlowOp (#15875)
• Fix: Initialize.inverse() now raises CircuitError immediately (#15782) (#15848)
• Fix: Initialize.inverse() now raises CircuitError immediately (#15782)

Changed

• Bumpy minimum NumPy version to 2.0 (#16243)
• Write policy on dependency versioning (#16239)

Deprecated

• Deprecations in qs_decomposition: opt_a* are not needed and decomposer_*q restrictions (#16158)
• Remove the commutation cache from the commutation checker (#15988)

Build System

• Use stestr to prune its own history in CI jobs (#16364)
• Add C-API slot-assignment compatibility check (#16301)
• Add slot-listing command to bindgen-cli (#16289)
• Rename bindgen-c to bindgen-cli (#16096)
• Add option to use mimalloc as global allocator (#16024)
• Bump MSRV to 1.87 and nalgebra to the 0.34 release (#16045)

Performance

• Update Sabre extended set on-the-fly (#14912)
• Use GateCount over Depth in Clifford+T pipeline (#16218)
• Move ParameterVector to Rust (#16228)
• Parallelize unitary synthesis (#16275)
• Use toposort instead of topological_op_nodes for DAG reconstruction (#15987)
• Perform ConsolidateBlocks analysis in parallel (#16230)
• Make Optimize1qGatesDecomposition multithreaded (#15567)
• Use np.bitwise_count in BitArray.bitcount (#15836)
• Use TwoQubitPeepholeOptimization in preset pass managers (#16136)
• Speedup SubstitutePi4Rotations (#16217)
• Parallelize the CommutationAnalysis pass (#16014)
• Pivot to foldhash for hash implementation in IndexMap (#15920)
• Update PauliProductRotationGate.to_matrix() from scipy to numpy (#16016)
• Add option to use mimalloc as global allocator (#16024)