v0.7.0

🚨 Breaking changes

• Renamed variables and attributes to make units consistent and explicit.
• Moved and renamed several submodules to make package organization more clear.
• ChemistryApplicationInstance has been refactored and renamed to MolecularHamiltonianGenerator.
• AutoCCZ and Litinski factories now both represented by the MagicStateFactory class to make it easier to swap them.
• Removed unused Taylorization LCU code.

🔥 Features

• Eliminated many warnings and updated tests to treat warnings as errors.
• Number of factories exposed as optional parameter for footprint analysis.
• Added breakdown of physical qubits into distillation vs computation for footprint analysis.
• Updated automatic resource estimator to take advantage of recent improvements to graph state compilation.

🐛 Fixes

• Litinski factory iterators use the physical error rate rather than architecture model type.
• Fixed bug in the number of logical qubits assumed for decoder estimates.
• Updated description of dependencies and examples in README.
• Fixed minor typing issues and added pyright as a automated check for some files.

v0.6.0

What's Changed

• Include examples for utility-scale estimates
• Add 3 new hamiltonians given by LANL application instances
• Fix the bottleneck of truncating space with FNO
• Increasing maximum data qubit code distance
• user accessible mlflow for scf tooling
• ability to optimize ruby slippers hyperparams
• remove isolated nodes before costing
• Convex Optimization to determine phases in QSP algorithm
• Implement Preparation of |COEF_int>
• Construct time marching circuit using the USVA
• Returning DF block encoding one-norm
• Added DetailedIonTrapHardwareModel
• Banish juliapkg
• Remove Python 3.11 upper bound and bump orquestra-* packages
• Update docstring for QuantumProgram and add various type hints

Full Changelog: v0.5.0...v0.6.0

v0.5.0

Features

• Installs stim by default
• Allows costing out double factorized Hamiltonians
• Implement performance testing on each PR merge
• Implement home grown footprint analysis
• New tools for linear differential equations: Invert block encodings and compression gadget
• Report proportion of quantum volume used for routing in ResourceInfo
• Faster GSC with new Ruby Sippers Compiler

Full Changelog: v0.4.0...v0.5.0

v0.4.0

What's Changed

• test: Add tests for graph compilation resource estimation by @max-radin in #13
• Changing hydrogen chain multiplicity and openfermion dependency by @max-radin in #14
• Update pyLIQTR dependency by @mstechly in #15
• Fix: Fix advanced example by @SebastianMorawiec in #17
• Make PySCF optional by @max-radin in #18
• Support AVAS in Hamiltonian generation by @max-radin in #16
• Add option to pass rotations through Jabalizer. by @mstechly in #20
• Feat: Remove deps on Pyquil due to conflicting deps with ray by @SebastianMorawiec in #19
• Removed using deprecated methods from pyLIQTR. by @mstechly in #21
• Changes to allow directly calling Jabalizer from Python by @ScottPJones in #23
• feat: Implement graph sim by @AthenaCaesura in #26
• Feat: Examples on how to use MS estimates on QE clusters by @SebastianMorawiec in #27
• Dta2 119/Refactor estimation pipeline by @dexter2206 in #25
• Create adjacency list and return with icm output by @ScottPJones in #28
• feat: write graph-sim in julia by @AthenaCaesura in #29
• Refactor transformers by @dexter2206 in #30
• Add the feature of reducing the virtual space with frozen natural orbitals (FNOs). by @mkodrycka in #22
• Replace graph compilation modules and update examples. by @mstechly in #31
• Add DecoderModel class. by @mstechly in #32
• Dta2 122/decoders update by @mstechly in #35
• Change naming of some methods/variables in GraphResourceEstimator by @dexter2206 in #36
• Dta2 98 fix how circuit gate synthesis accuracy is used by @akataba in #33
• Adjustments after call with Simon by @dexter2206 in #37
• Benchmarks for get_qsp_time_evolution_program by @dexter2206 in #38
• Add extrapolation pipeline by @AthenaCaesura in #39
• Dta 128/azure update by @mstechly in #40
• Update CODEOWNERS by @max-radin in #44
• Add SCF options to the ChemistryApplicationInstance class by @mkodrycka in #41
• Update README.md by @mstechly in #46
• Dta2 124 by @mstechly in #47
• Improve performance of get qsp program by @dexter2206 in #48
• Update examples by @mstechly in #49
• Fix incorrectly recognizing some XPowGates as Identities by @dexter2206 in #50
• Kj/fix benchmarks by @dexter2206 in #56
• adding my lcu project files by @ZPBQuantum in #8
• Dta2 145/example workflows by @mstechly in #57
• Enforcing SCF convergence by @max-radin in #45
• OpenFermion estimates without AVAS by @max-radin in #59
• Loosen pyscf version constraint by @max-radin in #60
• Update CODEOWNERS by @mstechly in #51
• Expose rank as parameter by @max-radin in #62
• DTA-2 132/Adjust expected results for commented out test cases by @dexter2206 in #64
• OpenFermion estimates using ERI by @max-radin in #63
• fix: extrapolation pipeline by @AthenaCaesura in #61
• Added the script for generating QAOA circuits for QUBO instances by @Guoming-Wang in #43
• fixup! adding my lcu project files (#8) by @mstechly in #58
• Fix: Fix orquestra example after recent changes by @SebastianMorawiec in #65
• Removing automatic resource estimator from top level package by @max-radin in #66
• Clean up ResourceInfo class by @dexter2206 in #69
• fix: small data path error by @1ethanhansen in #72
• Adding openfermion resource info by @max-radin in #70
• fix(architecture_model): variable names by @AthenaCaesura in #68
• feat: add estimators by @AthenaCaesura in #67
• Implement proportional distribution of synthesis failure tolerance by @dexter2206 in #71
• fix: implement qubit remapping so it takes into account all circuits at once by @dexter2206 in #75
• Implement Athena's ideas for reduced number of gate operations as well as optimize structures & performance by @ScottPJones in #73
• chore: unpin sdk version by @1ethanhansen in #76
• Update benchmarks by @dexter2206 in #80
• feat: small performance improvement by @AthenaCaesura in #79
• feat: new decoder model by @AthenaCaesura in #74
• fix(graph_estimator.py): underestimating logical cell failure rate by @AthenaCaesura in #82
• chore: bump orquestra quantum version by @1ethanhansen in #84
• fix(test_graph_sim_mini): ensure tableaus do not repeat stabilizers by @AthenaCaesura in #85
• Implement select_v decomposition by @dexter2206 in #81
• Fix toy problem by @AthenaCaesura in #83
• Implement new example workflow without pipelines by @dexter2206 in #86
• Low-depth GSEE utilities by @max-radin in #88
• Make BasicArchitectureModel a dataclass by @dexter2206 in #87
• feat: turn DecoderModel into a dataclass by @1ethanhansen in #89
• Add reset gate by @AthenaCaesura in #91
• Refactor HW models, widgets and how they are used in GraphEstimator by @dexter2206 in #90
• Feat: Enable graph data reuse in workflows by @SebastianMorawiec in #93
• ZQS-1334: Add COEF_k and PREP by @olgOk in #94
• feat: add ability to log benchq params to mlflow by @1ethanhansen in #95
• Double-factorized QPE by @max-radin in #96
• feat: add offset block encoding by @AthenaCaesura in #98
• Dta2 206 enable user to input surface code cycle time by @bdas4018 in #101
• fix(docs): example 5 by @1ethanhansen in #104
• docs: add new mlflow demo by @1ethanhansen in #102
• test: increase min coverage to 75 by @1ethanhansen in #99
• Feat: Workflow to release benchq to PyPi by @SebastianMorawiec in #107

Full Changelog: v0.3.0...v0.4.0

v0.3.0

• Updated resource estimation techniques
• Added Orquestra example
• Updated other examples
• bugfixes

v0.2.0

Full Changelog: https://github.com/zapatacomputing/benchq/commits/v0.2.0