Changelog

All notable changes to this project will be documented in this file.

The format is based on a mixture of Keep a Changelog and Common Changelog. This project adheres to Semantic Versioning, with the exception that minor releases may include breaking changes.

Unreleased

Added

General MQT Compiler Collection infrastructure

• ✨ Launch the MQT Compiler Collection with the QC and QCO dialects, its core compiler infrastructure, and C++ and Python APIs (#1264, #1330, #1402, #1428, #1430, #1436, #1443, #1446, #1464, #1465, #1470, #1471, #1472, #1474, #1475, #1506, #1510, #1513, #1521, #1542, #1548, #1550, #1554, #1567, #1569, #1570, #1572, #1573, #1580, #1602, #1603, #1620, #1623, #1626, #1627, #1635, #1638, #1673, #1675, #1700, #1717, #1728, #1730, #1749, #1751, #1762, #1765, #1780, #1781, #1782, #1806, #1807, #1808, #1815, #1824, #1869, #1872, #1914, #1925, #1927, #1935, #1936, #1938, #1975, #1976, #2006, #2014, #2015, #2017, #2026, #2028, #2054, #2058, #2125, #2136, #2149, #2150, #2158, #2194, #2210, #2211, #2215, #2218, #2220, #2323, #2336, #2344) (@burgholzer, @denialhaag, @taminob, @DRovara, @li-mingbao, @Ectras, @MatthiasReumann, @simon1hofmann, @J4MMlE)

• ✨ Add a library for typed structured quantum benchmarks with versioned instance specifications, analytic references, deterministic manifests, and C++, Python, and command-line interfaces (#2135, #2299, #2315, #2324, #2337, #2380, #2402, #2404, #2409, #2410, #2493) (@burgholzer, @denialhaag)

• ✨ Add DD construction, simulation, statevector extraction, and sampling for QCO programs with structured control and dynamic quantum data, including direct lowering and dense-array helpers for supported compiler inputs (#1915, #1973, #2077, #2078, #2079, #2334) (@simon1hofmann, @burgholzer)

• ✨ Add immutable MLIR compiler targets, selected payload specifications, payload-aware control-flow legalization, QDMI device integration, ordered operation applicability, directional native synthesis, and target compilation through C++, Python, and mqt-cc (#2285, #2219, #2162, #2049, #1999, #1993, #1687) (@MatthiasReumann, @simon1hofmann, @burgholzer)

Import and export

• ✨ Add Qiskit circuit import, target-aware export, and reusable custom Gate round trips to the compiler collection (#2031, #2133, #2140, #2150, #2175, #2176, #2178, #2342) (@burgholzer, @simon1hofmann)

• ✨ Add conversions between jeff and QCO (#1479, #1548, #1565, #1637, #1676, #1706, #1776, #1836, #1934, #2000, #2018, #2105, #2339, #2457) (@denialhaag, @burgholzer)

• ✨ Add QIR generation support to the MQT Compiler Collection (#1264, #1446, #1513, #1521, #1548, #1567, #1569, #1570, #1572, #1580, #1620, #1624, #1626, #1648, #1710, #1751, #1755, #1787, #1815, #1823, #1933, #1978, #1979, #2007, #2026, #2030, #2066, #2217, #2340) (@burgholzer, @denialhaag, @simon1hofmann, @li-mingbao, @DRovara, @MatthiasReumann)

• ✨ Add OpenQASM import and export to the MQT Compiler Collection, including fixed-angle constants and proven affine quantum-register indices (#1910, #1987, #1994, #2003, #2026, #2169, #2203, #2338) (@burgholzer, @denialhaag)

Passes and transformations

• ✨ Add passes for quantum-specific interprocedural optimizations (#2193) (@DRovara, @burgholzer)

• ✨ Add Pauli twirling, quantum loop unrolling, and qubit reuse passes (#1705, #1718, #1755, #1756, #1923, #1924, #2039, #2118, #2216, #2224) (@MatthiasReumann, @DRovara, @burgholzer, @simon1hofmann)

• ✨ Add a compiler-target-aware place-and-route pass (#1537, #1547, #1568, #1581, #1583, #1588, #1600, #1664, #1709, #1716, #1748, #1805, #1870, #1904, #1911, #1951, #1956, #1997, #2016, #2060, #2179, #2184, #2185, #2205, #2240, #2436) (@MatthiasReumann, @burgholzer, @rturrado, @simon1hofmann)

• ✨ Add modifier and global-phase normalization passes (#1986, #1995, #2015) (@burgholzer, @denialhaag)

• ✨ Add single-qubit optimization passes for unitary fusion, Hadamard lifting, and rotation merging (#1407, #1605, #1672, #1674, #2002, #2038, #2228, #2478) (@J4MMlE, @lirem101, @burgholzer, @denialhaag, @MatthiasReumann, @simon1hofmann)

• ✨ Add multi-qubit decomposition, fusion, and target-native synthesis passes (#1774, #1802, #1803, #1809, #1810, #1814, #1832, #1850, #1865, #1961, #1996, #1998, #2001, #2467, #2468, #2478) (@simon1hofmann, @burgholzer)

Other additions

• ✨ Support DDSIM QDMI statevector extraction for Adaptive Profile QIR with classical control flow, dynamic allocation, and terminal measurements.

• ✨ Expose ordered shots from DDSIM QDMI QIR jobs, with matching histograms (#2368) (@burgholzer)

• 🐳 Add dev container configuration for a consistent local development environment (#1786) (@denialhaag)

Changed

• 💥 Require CMake 3.28 or newer for source builds and embedded projects, use native dependency exclusions and system include handling, and disable unused C++ module scanning (#2421) (@burgholzer)

• ⬆️ Update clang-tidy to version 23 and adapt the C++ sources to its diagnostics (#2328) (@simon1hofmann)

• 💥 Update QIR execution for QIR 2.1, isolated runtimes, deterministic QDMI sampling, and safe statevector extraction (#2035, #2036, #2246) (@burgholzer, @denialhaag)

• 🔧 Build the LLVM/MLIR 23.1 compiler infrastructure and DDSIM QIR support by default, with one build option for source builds that omit both parts (#1356, #1549, #1953, #2284, #2298) (@burgholzer, @denialhaag, @simon1hofmann)

Removed

• 💥 Remove qc::QuantumComputation, MQT::CoreIR, and MQT::CoreQASM. Use compiler-backed QC/QCO APIs in v4 or the MQT Core v3 release series for legacy code (#2111, #2112, #2288) (@burgholzer, @simon1hofmann)

• 💥 Remove the standalone QIR runner and make the QIR runtime and JIT internal DDSIM implementation details (#2246) (@denialhaag)

3.10.0 - 2026-09-05

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Expose ordered shots from DDSIM QDMI OpenQASM jobs, with matching histograms (#2368) (@burgholzer)

Changed

• ⚡ Run PennyLane QDMI jobs concurrently and release the GIL during waits and result retrieval (#2349) (@burgholzer)

• 💥 Raise the minimum Qiskit version from 1.1.0 to 2.1.0 (#2358) (@burgholzer)

• 💥 Replace the QDMI-specific primitives with native Qiskit primitives and typed backend factories. Sampler and memory=True require genuine QDMI SHOTS (#2358) (@burgholzer)

• ⬆️ Update nanobind to version 3.0.1 (#2209, #2283) (@denialhaag, @burgholzer)

• 💥 Move circuit IR OpenQASM serialization from operation subclasses to qasm3::Serializer in qasm3/Serializer.hpp (#2249) (@simon1hofmann)

• 💥 Drop support for x86 macOS and stop publishing the respective wheels (#2259) (@denialhaag)

• ⬆️ Raise the macOS deployment target to 13.3 to enable std::format in libc++ (#2259) (@denialhaag)

• 💥 Require Python 3.11 or newer (#2209) (@denialhaag, @burgholzer)

• 📦 Publish one split-mode cp311-abi3 wheel for GIL-enabled CPython 3.11 and newer (#2209) (@denialhaag, @burgholzer)

• 📦 Publish one cp315-abi3t wheel for free-threaded CPython 3.15 and newer (#2209) (@denialhaag, @burgholzer)

• ⚡ Remove an extra dense copy from VectorDD.get_vector (#2209) (@burgholzer)

• 🐛 Protect process-wide DD, IR, and QDMI state for free-threaded Python (#2209) (@burgholzer)

• 💥 Prune dead and misleading CoreIR APIs, including renaming the non-garbage logical output count to getNoutputQubits() and num_output_qubits (#2112) (@simon1hofmann)

Removed

• 💥 Remove test-only DD state generators, recursive functionality construction, and DD-specific named-gate helpers (#2257, #2335) (@simon1hofmann)

• 💥 Remove the spdlog dependency from source builds, installed CMake packages, and Python wheels. QDMI diagnostics continue to be written to standard error (#2270) (@denialhaag)

• 💥 Remove CircuitOptimizer. Move circuit flattening and final-measurement removal to QuantumComputation, equivalence-checking transformations to MQT QCEC, and mapping transformations to MQT QMAP. Move single-qubit gate fusion to both downstream packages. Remove the public circuit dependency graph and transformations without production consumers (#2262) (@simon1hofmann)

• 💥 Remove MQT::CoreAlgorithms, its fixed-circuit factories, and the legacy DD package evaluation. MQT Core provides no direct replacement (#2214) (@burgholzer)

• 💥 Remove the unused decision-diagram approximation algorithm, including the dd/Approximation.hpp header, dd::ApproximationMetadata, and dd::approximate. No replacement is provided (#2154) (@burgholzer)

• 💥 Remove nlohmann_json from the public package contract. MQT Core no longer installs or exports the library, no installed header exposes a nlohmann type, and the decision-diagram statistics report through strings and streams (#2138) (@denialhaag)

• 💥 Remove the neutral-atom stack, which moves to MQT QMAP. This drops the neutral-atom computation model, the neutral-atom FoMaC device session, the neutral-atom QDMI device and its configuration, the mqt.core.na Python module, AodOperation, and the Move, Bridge, AodActivate, AodDeactivate, and AodMove operation kinds (#2137) (@denialhaag)

• 💥 Remove the random-number generator, seed, and getGenerator() method from QuantumComputation; randomized algorithms now own generators initialized from their seed arguments (#2111) (@simon1hofmann)

• 💥 Remove the FoMaC compatibility name from the C++ and Python QDMI APIs. Use the qdmi C++ namespace, headers, library, and CMake target; the mqt.core.qdmi Python module; and module-level functions in mqt.core.qdmi.driver (#2115) (@burgholzer)

• 💥 Remove the ZX-calculus library, including the mqt-core-zx target, MQT::CoreZX alias, zx headers and namespace, and its Boost.Multiprecision and GMP build support. Equivalence-checking users should use MQT QCEC; its ZX implementation is internal and does not provide a replacement public API (#2082) (@burgholzer)

• 🔥 Remove density matrix support from the DD package (#1466) (@burgholzer)

• 🔥 Remove datastructures (ds) (sub)library (#1458) (@burgholzer)

Fixed

• 🐛 Initialize Qiskit classical bits before OpenQASM 3 serialization so partially measured circuits preserve their zero values (#2399) (@burgholzer)

• 🐛 Handle empty DDSIM results and NUL-terminated QDMI result buffers (#2288) (@simon1hofmann)

• 🐛 Validate output permutations before I/O mapping initialization (#2278) (@denialhaag)

3.9.2 - 2026-08-26

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Allow C++ and Python QDMI job submissions to omit the shot count, leaving repetition semantics to the program and device (#2258) (@burgholzer)

3.9.1 - 2026-08-25

If you are upgrading: please see UPGRADING.md.

Added

• 🚸 Let PennyLane QDMI devices reuse an already-open session, including a device selected from a Slurm license (#2232) (@burgholzer)

• ✨ Let a package register a program serializer for a program format through the mqt.core.qiskit.program_serializers entry point group (#2114) (@marcelwa, @burgholzer)

• ✨ Add mqt.core.qdmi.is_binary_program_format, which states whether a program format requires exact-byte submission (#2114) (@marcelwa, @burgholzer)

Removed

• 💥 Remove the IQM JSON converter qiskit_to_iqm_json and the MoveGate from the Qiskit plugin, which QDMI-on-IQM now owns (#2114) (@marcelwa, @burgholzer)

3.9.0 - 2026-08-19

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Add PennyLane support for gate-based QDMI devices (#2005, #2147) (@burgholzer, @marcelwa)

• ✨ Add Device::submitCalibrationJob and mqt.core.qdmi.Device.submit_calibration_job for triggering a calibration run (#2148) (@marcelwa, @burgholzer)

• ✨ Add SpecAudits, a method and probe script for auditing tests that pin behavior the project never specified (#2124) (@marcelwa)

• 🚸 Add typed stable-ID construction for Qiskit backends, lazy provider discovery, and sampler and estimator factories with explicit shot and precision defaults (#2084) (@burgholzer)

• 🧪 Test static Slurm license admission and QDMI execution with DDSIM and the superconducting device, and document the cluster setup (#2043) (@burgholzer)

• ✨ Add C++ FoMaC and Python QDMI adapters that open the device named by one local Slurm license environment value (#2025) (@burgholzer)

• ✨ Add generic C++ FoMaC and Python QDMI support for custom device properties that contain operation handles (#2042) (@burgholzer)

• 📝 Generate llms.txt documentation indexes with Sphinx-LLM (#1989, #2046) (@denialhaag, @burgholzer)

• ✨ Support retrieving existing jobs by ID through the QDMI client API, C++ FoMaC API, and Python QDMI API, and expose optional device queue length and job queue position (#2008, #2010) (@burgholzer)

• 🐍 Build CPython 3.15 wheels. Their post-build tests remain disabled until test dependency wheels are available (#2011) (@denialhaag)

• ✨ Bundle reusable IQM Garnet and Emerald superconducting device models with stable QDMI registry IDs (#1992) (@burgholzer)

• ✨ Expose compressed vector and matrix DD serialization through bytes-based Python APIs (#1983) (@burgholzer)

• ✨ Make the neutral-atom and superconducting QDMI devices runtime configurable with session-owned topology, operations, and calibration data (#1974, #1980) (@burgholzer)

• ✨ Expose registered QDMI device IDs without loading device libraries (#1972) (@burgholzer)

• ✨ Add typed runtime configuration transport and relocatable assets for QDMI device descriptions (#1967) (@burgholzer)

Changed

• ⬆️ Update QDMI to version 1.3.3 (#2168) (@denialhaag)

• ⬆️ Update nanobind to version 2.15.0 (#2141) (@denialhaag)

• 💥 Replace the MQT-specific QDMI primitive options mappings with explicit shot and precision defaults (#2084) (@burgholzer)

• ♻️ Simplify Python optional-dependency checks while preserving the Qiskit and PennyLane availability flags (#2108) (@simon1hofmann)

• 💥 Remove the unused pybind11 CMake helper and rename add_mqt_python_binding_nanobind to add_mqt_python_binding (#2106) (@denialhaag)

• 💥 Move Python QDMI entities and the neutral-atom specialization to QDMI namespaces, expose device registration and opening through mqt.core.qdmi.driver, retain v3 FoMaC compatibility aliases, and let the Qiskit adapter open stable device IDs directly (#2074) (@burgholzer)

• 🚀 Reduce ZX diagram growth for multi-controlled X gates with an exact ancilla-free quadratic decomposition (#1984) (@burgholzer)

Fixed

• 🐛 Distinguish scalar OpenQASM qubits from one-element qubit registers and reject indexing scalar qubits (#2157) (@DRovara, @burgholzer)

• 🐛 Preserve the original OpenQASM type error when an assignment’s right-hand expression cannot be typed (#2156) (@DRovara, @burgholzer)

Removed

• 💥 Remove batch job submission from the QDMI client. Device::submitJob now states that MQT Core does not support batch jobs (#2148) (@marcelwa, @burgholzer)

• 💥 Remove QDMI device configuration through [tool.qdmi] in pyproject.toml and the vendored toml++ header (#2116) (@denialhaag)

3.8.0 - 2026-07-30

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Add binary-safe QDMI program submission and retrieval to FoMaC, including explicit text and exact-byte APIs and all standard QDMI program formats (#1957) (@burgholzer)

• ✨ Add versioned, relocatable configuration and stable-ID registration for QDMI device libraries, including disabled-ID reservations, fresh device sessions, idempotent registration, and external-device target metadata (#1912) (@burgholzer)

• ✨ Add native relative-phase CCX (rccx) support across the IR, DD package, ZX diagrams, OpenQASM import/export, and Python/Qiskit bindings (#1886, #1950) (@simon1hofmann)

• ✨ Add support for QDMI child devices to the driver and FoMaC libraries (#1897, #1952) (@burgholzer)

• ✨ Add typed custom property and result queries to the C++ and Python FoMaC libraries (#1895) (@burgholzer)

• ✨ Add support for custom job parameters to C++ and Python FoMaC library (#1887) (@flowerthrower, @burgholzer)

• ✨ Add labeled and ordered output schemas to the QIR runtime (#1877) (@rturrado)

• ✨ Add boolean, integer, floating-point, tuple, and array record output functions to the QIR runtime (#1799) (@rturrado)

• ✨ Add the reusable in-process MQT::CoreQIRJIT library and QIR program format support to the DDSIM QDMI device (#1766) (@rturrado)

Changed

• ⬆️ Raise the minimum supported QDMI version to 1.3.2 (#1897) (@burgholzer)

Removed

• 🔥 Replace the unstable C++ Driver::addDynamicDeviceLibrary and Python add_dynamic_device_library APIs with definition registration and stable-ID opening (#1912) (@burgholzer)

Fixed

• 🐛 Allow MQT Core to be embedded as a CMake subproject without target collisions and make its bundled QDMI devices individually configurable (#1965) (@burgholzer)

• 🐛 Fix QIR function names for adjoint gates (#1830) (@denialhaag)

3.7.0 - 2026-07-09

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Add support for IQM’s move gate in the QDMI Qiskit backend converter (#1844, #1848) (@burgholzer, @marcelwa)

• 🚸 Add const version of the CompoundOperation’s getOps() function (#1826) (@ystade)

• 🚸 Add CMake presets to provide a standardized and reproducible way to configure builds (#1660) (@denialhaag)

Changed

• ⬆️ Update QDMI to version 1.3.2 (#1873) (@denialhaag)

• ♻️ Improve implementation and usability of FoMaC classes (#1849) (@MatthiasReumann)

• ⬆️ Update nanobind to version 2.13.0 (#1817)

• ⬆️ Update munich-quantum-toolkit/workflows to version v2.0.1 (#1660, #1737) (@denialhaag)

Removed

• 📝 Remove support for generating LaTeX documentation (#1828) (@denialhaag)

Fixed

• 🐛 Fix invalid prop_type for QDMI_DEVICE_PROPERTY_COUPLINGMAP in QDMI SC Device (#1842) (@MatthiasReumann)

3.6.1 - 2026-05-20

Changed

• 🚸 Improve native gate support for the Qiskit-to-OpenQASM3 conversion in the QDMI-Qiskit interface (#1719) (@burgholzer)

Fixed

• 🏁 Fix dynamic loading of QDMI device DLLs on Windows when an absolute path is provided (#1720) (@burgholzer)

3.6.0 - 2026-05-13

If you are upgrading: please see UPGRADING.md.

Added

• 🚸 Add a measurement instruction to the default SC QDMI device (#1694) (@burgholzer)

• ✨ Add support for multi-controlled gates to the QDMI Qiskit backend converter (#1694) (@burgholzer)

Changed

• ♻️ Build all built-in QDMI devices as shared libraries (#1694) (@burgholzer)

• ⬆️ Update minimum supported Qiskit version to 1.1.0 (#1694) (@burgholzer)

Fixed

• 🐛 Fix missing nlohmann_json.natvis in Windows component-based CMake installs (#1702) (@burgholzer)

• 🐛 Fix segfault in DD sample method when idle classical bits are present (#1694) (@burgholzer)

Removed

• 🔥 Remove shared library wrappers for QDMI devices (#1694) (@burgholzer)

3.5.1 - 2026-04-23

If you are upgrading: please see UPGRADING.md.

Fixed

• 🐛 Fix malformed include directories in exported nlohmann_json CMake targets for component-based installs (#1662) (@burgholzer)

3.5.0 - 2026-04-21

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Add support for multi-controlled gates to ZX package (#1380) (@keefehuang, @denialhaag)

• ✨ Add Sampler and Estimator primitives to the QDMI-Qiskit interface (#1507) (@marcelwa)

Changed

• ⬆️ Update nanobind to version 2.12.0 (#1528)

• ⬆️ Update QDMI to version 1.3.0 (#1652) (@burgholzer)

• 📦 Switch to component-based installation for the MQT Core Python package (#1596) (@burgholzer)

• ⬆️ Update QDMI to latest version from stable v1.2.x branch (#1593) (@burgholzer)

• ⬆️ Update clang-tidy to version 22 (#1564) (@denialhaag, @burgholzer)

• 👷 Build on macos-26/macos-26-intel by default and macos-15/macos-15-intel for extensive tests (#1571) (@denialhaag)

3.4.1 - 2026-02-01

Changed

• ⬆️ Update nanobind to version 2.11.0 (#1481) (@denialhaag)

• ⬆️ Update Boost to version 1.89.0 (#1453) (@burgholzer)

• ⬆️ Update QDMI to latest version from stable v1.2.x branch (#1453) (@burgholzer)

• ⬆️ Update spdlog to version 1.17.0 (#1453) (@burgholzer)

• ♻️ Use llc instead of random clang for compiling QIR test circuits to improve robustness and handle opaque pointers correctly across LLVM versions (#1447) (@burgholzer)

• ♻️ Extract singleton pattern into reusable template base class for QDMI devices and driver (#1444) (@ystade, @burgholzer)

• 🚚 Reorganize QDMI code structure by moving devices into dedicated subdirectories and separating driver and common utilities (#1444) (@ystade)

Removed

• 🔥 No longer actively type check Python code with mypy and solely rely on ty (#1437) (@burgholzer)

3.4.0 - 2026-01-08

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Return device handle from add_dynamic_device_library for direct backend creation (#1381) (@marcelwa)

• ✨ Add IQM JSON support for job submission in Qiskit-QDMI Backend (#1375, #1382) (@marcelwa, @burgholzer)

• ✨ Add authentication support for QDMI sessions with token, username/password, auth file, auth URL, and project ID parameters (#1355) (@marcelwa)

• ✨ Add a new QDMI device that represents a superconducting architecture featuring a coupling map (#1328) (@ystade)

• ✨ Add bi-directional iterator that traverses the def-use chain of a qubit value (#1310) (@MatthiasReumann)

• ✨ Add OptionalDependencyTester to lazily handle optional Python dependencies like Qiskit (#1243) (@marcelwa, @burgholzer)

• ✨ Expose the QDMI job interface through FoMaC (#1243) (@marcelwa, @burgholzer)

• ✨ Add Qiskit backend wrapper with job submission support for QDMI devices through a provider interface (#1243, #1385) (@marcelwa, @burgholzer)

• ✨ Support QDMI_DEVICE_PROPERTY_SUPPORTEDPROGRAMFORMATS in the NA QDMI Device and the DDSIM QDMI Device (#1243) (@marcelwa, @burgholzer)

• ✨ Support QDMI_DEVICE_JOB_PROPERTY_PROGRAM in the NA QDMI Device (#1243) (@marcelwa, @burgholzer)

Changed

• 📦🏁 Build Windows x86 wheels on windows-2025 runner for newer compiler (#1415) (@burgholzer)

• 👷 Build on macos-15/windows-2025 by default and macos-14/windows-2022 for extensive tests (#1414) (@burgholzer)

• 📦🍎 Build macOS arm64 wheels on macos-15 runner for newer compiler (#1413) (@burgholzer)

• ⚡ Improve uv build caching by removing unconditional reinstall-package and configuring dedicated cache-keys (#1412) (@burgholzer)

• 👨‍💻📦 Build spdlog and QDMI generators as shared libraries in Python package builds (#1411, #1403) (@burgholzer)

• ♻️🏁 Remove Windows-specific restrictions for dynamic QDMI device library handling (#1406) (@burgholzer)

• ♻️ Migrate Python bindings from pybind11 to nanobind (#1383) (@denialhaag, @burgholzer)

• 📦️ Provide Stable ABI wheels for Python 3.12+ (#1383) (@burgholzer, @denialhaag)

• 🚚 Create dedicated mqt.core.na submodule to closely follow the structure of other submodules (#1383) (@burgholzer)

• ✨ Add common definitions and utilities for QDMI (#1355) (@burgholzer)

• 🚚 Move NA QDMI device in its right place next to other QDMI devices (#1355) (@burgholzer)

• ♻️ Allow repeated loading of QDMI device library with potentially different session configurations (#1355) (@burgholzer)

• ♻️ Enable thread-safe reference counting for QDMI devices singletons (#1355) (@burgholzer)

• ♻️ Refactor FoMaC singleton to instantiable Session class with configurable authentication parameters (#1355) (@marcelwa)

• 👷 Stop testing on ubuntu-22.04 and ubuntu-22.04-arm runners (#1359) (@denialhaag, @burgholzer)

• 👷 Stop testing with clang-19 and start testing with clang-21 (#1359) (@denialhaag, @burgholzer)

• 👷 Fix macOS tests with Homebrew Clang via new munich-quantum-toolkit/workflows version (#1359) (@denialhaag, @burgholzer)

• 👷 Re-enable macOS tests with GCC by disabling module scanning (#1359) (@denialhaag, @burgholzer)

• ♻️ Group circuit operations into scheduling units for MLIR routing (#1301) (@MatthiasReumann)

• 👷 Use munich-quantum-software/setup-mlir to set up MLIR (#1294) (@denialhaag)

• ♻️ Preserve tuple structure and improve site type clarity of the MQT NA Default QDMI Device (#1299) (@marcelwa)

• ♻️ Move DD package evaluation module to standalone script (#1327) (@burgholzer)

• ⬆️ Bump QDMI version to 1.2.0 (#1243) (@marcelwa, @burgholzer)

Fixed

• 🔧 Install all available QDMI device targets in Python package builds (#1403) (@burgholzer)

• 🐛 Fix operation validation in Qiskit backend to handle device-specific gate naming conventions (#1384) (@marcelwa)

• 🐛 Fix conditional branch handling when importing MLIR from QuantumComputation (#1378) (@lirem101)

• 🐛 Fix custom QDMI property and parameter handling in SC and NA devices (#1355) (@burgholzer)

• 🚨 Fix argument naming of QuantumComputation and CompoundOperation dunder methods for properly implementing the MutableSequence protocol (#1338) (@burgholzer)

• 🐛 Fix memory management in dynamic QDMI device by making it explicit (#1336) (@ystade)

Removed

• 🔥 Remove wheel builds for Python 3.13t (#1371) (@burgholzer)

• 🔥 Remove the evaluation extra from the MQT Core Python package (#1327) (@burgholzer)

• 🔥 Remove the mqt-core-dd-compare entry point from the MQT Core Python package (#1327) (@burgholzer)

3.3.3 - 2025-11-10

Added

• ✨ Add support for bridge gates for the neutral atom hybrid mapper (#1293) (@lsschmid)

Fixed

• 🐛 Revert change to opTypeFromString() signature made in #1283 (#1300) (@denialhaag)

3.3.2 - 2025-11-04

Added

• ✨ Add DD-based simulator QDMI device (#1287) (@burgholzer)

• ✨ A --reuse-qubits pass implementing an advanced form of qubit reuse to reduce the qubit count of quantum circuits (#1108) (@DRovara)

• ✨ A --lift-measurements pass that attempts to move measurements up as much as possible, used for instance to enable better qubit reuse (#1108) (@DRovara)

• ✨ Add native support for R(theta, phi) gate (#1283) (@burgholzer)

• ✨ Add A*-search-based routing algorithm to MLIR transpilation routines (#1237, #1271, #1279) (@MatthiasReumann)

Fixed

• 🐛 Fix edge-case in validation of NAComputation (#1276) (@ystade)

• 🐛 Allow integer QASM version declarations (#1269) (@denialhaag)

3.3.1 - 2025-10-14

Fixed

• 🐛 Ensure spdlog dependency can be found from mqt-core install (#1263) (@burgholzer)

3.3.0 - 2025-10-13

If you are upgrading: please see UPGRADING.md.

Added

• 👷 Enable testing on Python 3.14 (#1246) (@denialhaag)

• ✨ Add dedicated PlacementPass to MLIR transpilation routines (#1232) (@MatthiasReumann)

• ✨ Add an NA-specific FoMaC implementation (#1223, #1236) (@ystade, @burgholzer)

• ✨ Enable import of BarrierOp into MQTRef (#1224) (@denialhaag)

• ✨ Add naive quantum program routing MLIR pass (#1148) (@MatthiasReumann)

• ✨ Add QIR runtime using DD-based simulation (#1210) (@ystade, @burgholzer)

• ✨ Add SWAP reconstruction patterns to the newly-named SwapReconstructionAndElision MLIR pass (#1207) (@taminob, @burgholzer)

• ✨ Add two-way conversions between MQTRef and QIR (#1091) (@li-mingbao)

• 🚸 Define custom assembly formats for MLIR operations (#1209) (@denialhaag)

• ✨ Add support for translating IfElseOperations to the MQTRef MLIR dialect (#1164) (@denialhaag, @burgholzer)

• ✨ Add MQT’s implementation of a generic FoMaC with Python bindings (#1150, #1186, #1223) (@ystade)

• ✨ Add new MLIR pass ElidePermutations for SWAP gate elimination (#1151) (@taminob)

• ✨ Add new pattern to MLIR pass GateElimination for identity gate removal (#1140) (@taminob)

• ✨ Add Clifford block collection pass to CircuitOptimizer module (#885) ([jannikpflieger], @burgholzer)

• ✨ Add isControlled() method to the UnitaryInterface MLIR class (#1157) (@taminob, @burgholzer)

• 📝 Integrate generated MLIR documentation (#1147) (@denialhaag, @burgholzer)

• ✨ Add IfElseOperation to C++ library and Python package to support Qiskit’s IfElseOp (#1117) (@denialhaag, @burgholzer, @lavanya-m-k)

• ✨ Add allocQubit and deallocQubit operations for dynamically working with single qubits to the MLIR dialects (#1139) (@DRovara, @burgholzer)

• ✨ Add qubit operation for static qubit addressing to the MLIR dialects (#1098, #1116) (@MatthiasReumann)

• ✨ Add MQT’s implementation of a QDMI Driver (#1010) (@ystade)

• ✨ Add MQT’s implementation of a QDMI Device for neutral atom-based quantum computing (#996, #1010, #1100) (@ystade, @burgholzer)

• ✨ Add translation from QuantumComputation to the MQTRef MLIR dialect (#1099) (@denialhaag, @burgholzer)

• ✨ Add reset operations to the MLIR dialects (#1106) (@DRovara)

Changed

• ♻️ Replace custom AllocOp, DeallocOp, ExtractOp, and InsertOp with MLIR-native memref operations (#1211) (@denialhaag)

• 🚚 Rename MLIR pass ElidePermutations to SwapReconstructionAndElision (#1207) (@taminob)

• ⬆️ Require LLVM 21 for building the MLIR library (#1180) (@denialhaag)

• ⬆️ Update to version 21 of clang-tidy (#1180) (@denialhaag)

• 🚚 Rename MLIR pass CancelConsecutiveInverses to GateElimination (#1140) (@taminob)

• 🚚 Rename xxminusyy to xx_minus_yy and xxplusyy to xx_plus_yy in MLIR dialects (#1071) (@BertiFlorea, @denialhaag)

• 🚸 Add custom assembly format for operations in the MLIR dialects (#1139) (@burgholzer)

• 🚸 Enable InferTypeOpInterface in the MLIR dialects to reduce explicit type information (#1139) (@burgholzer)

• 🚚 Rename check-quantum-opt test target to mqt-core-mlir-lit-test (#1139) (@burgholzer)

• ♻️ Update the measure operations in the MLIR dialects to no longer support more than one qubit being measured at once (#1106) (@DRovara)

• 🚚 Rename XXminusYY to XXminusYYOp and XXplusYY to XXplusYYOp in MLIR dialects (#1099) (@denialhaag)

• 🚚 Rename MQTDyn MLIR dialect to MQTRef (#1098) (@MatthiasReumann)

Removed

• 🔥 Drop support for Python 3.9 (#1181) (@denialhaag)

• 🔥 Remove ClassicControlledOperation from C++ library and Python package (#1117) (@denialhaag)

Fixed

• 🐛 Fix CMake installation to make find_package(mqt-core CONFIG) succeed (#1247) (@burgholzer, @denialhaag)

• 🏁 Fix stack overflows in OpenQASM layout parsing on Windows for large circuits (#1235) (@burgholzer)

• ✨ Add missing StandardOperation conversions in MLIR roundtrip pass (#1071) (@BertiFlorea, @denialhaag)

3.2.1 - 2025-08-01

Fixed

• 🐛 Fix usage of std::accumulate by changing accumulator parameter from reference to value (#1089) (@denialhaag)

• 🐛 Fix erroneous contains check in DD package (#1088) (@denialhaag)

3.2.0 - 2025-07-31

If you are upgrading: please see UPGRADING.md.

Added

• 🐍 Start building CPython 3.14 wheels (#1076) (@denialhaag)

• ✨ Add MQT-internal MLIR dialect conversions (#1001) (@li-mingbao)

Changed

• ✨ Expose enums to Python via pybind11’s new (enum.Enum-compatible) py::native_enum (#1075) (@denialhaag)

• ⬆️ Require C++20 (#897) (@burgholzer, @denialhaag)

3.1.0 - 2025-07-11

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Add MLIR pass for merging rotation gates (#1019) (@denialhaag)

• ✨ Add functions to generate random vector DDs (#975) (@MatthiasReumann)

• ✨ Add function to approximate decision diagrams (#908) (@MatthiasReumann)

• 📦 Add Windows ARM64 wheels (#926) (@burgholzer)

• 📝 Add documentation page for MLIR (#931) (@ystade)

• ✨ Initial implementation of the mqtdyn Dialect (#900) (@DRovara, @ystade)

Fixed

• 🐛 Fix bug in MLIR roundtrip passes caused by accessing an invalidated iterator after erasure in a loop (#932) (@flowerthrower)

• 🐛 Add missing support for sxdg gates in Qiskit circuit import (#930) (@burgholzer)

• 🐛 Fix bug related to initialization of operations with duplicate operands (#964) (@ystade)

• 🐛 Open issue for Qiskit upstream test only when the test is actually failing not when it was cancelled (#973) (@ystade)

• 🐛 Fix parsing of GPhase in the MQTOpt MLIR dialect (#1042) (@ystade, @DRovara)

Changed

• ⬆️ Bump shared library ABI version from 3.0 to 3.1 (#1047) (@denialhaag)

• ♻️ Switch from reference counting to mark-and-sweep garbage collection in decision diagram package (#1020) (@MatthiasReumann, [burgholzer], q-inho)

• ♻️ Move the C++ code for the Python bindings to the top-level bindings directory (#982) (@denialhaag)

• ♻️ Move all Python code (no tests) to the top-level python directory (#982) (@denialhaag)

• ⚡ Improve performance of getNqubits for StandardOperations (#959) (@ystade)

• ♻️ Move Make-State Functionality To StateGeneration (#984) (@MatthiasReumann)

• ♻️ Outsource definition of standard operations from MLIR dialects to reduce redundancy (#933) (@ystade)

• ♻️ Unify operands and results in MLIR dialects (#931) (@ystade)

• ⏪️ Restore support for (MLIR and) LLVM v19 (#934) (@flowerthrower, @ystade)

• ⬆️ Update nlohmann_json to v3.12.0 (#921) (@burgholzer)

3.0.2 - 2025-04-07

Added

• 📝 Add JOSS journal reference and citation information (#913) (@burgholzer)

• 📝 Add new links to Python package metadata (#911) (@burgholzer)

Fixed

• 📝 Fix old links in Python package metadata (#911) (@burgholzer)

3.0.1 - 2025-04-07

Fixed

• 🐛 Fix doxygen build on RtD to include C++ API docs (#912) (@burgholzer)

3.0.0 - 2025-04-06

If you are upgrading: please see UPGRADING.md.

Added

• ✨ Ship shared C++ libraries with mqt-core Python package (#662) (@burgholzer)

• ✨ Add Python bindings for the DD package (#838) (@burgholzer)

• ✨ Add direct MQT QuantumComputation to Qiskit QuantumCircuit export (#859) (@burgholzer)

• ✨ Support for Qiskit 2.0+ (#860) (@burgholzer)

• ✨ Add initial infrastructure for MLIR within the MQT (#878, #879, #892, #893, #895) (@burgholzer, @ystade, @DRovara, @flowerthrower, @BertiFlorea)

• ✨ Add State Preparation Algorithm (#543) (@M-J-Hochreiter)

• 🚸 Add support for indexed identifiers to OpenQASM 3 parser (#832) (@burgholzer)

• 🚸 Allow indexed registers as operation arguments (#839) (@burgholzer)

• 📝 Add documentation for the DD package (#831) (@burgholzer)

• 📝 Add documentation for the ZX package (#817) (@pehamTom)

• 📝 Add C++ API docs setup (#817) (@pehamTom, @burgholzer)

Changed

• Breaking: 🚚 MQT Core has moved to the munich-quantum-toolkit GitHub organization

• Breaking: ✨ Adopt PEP 735 dependency groups (#762) (@burgholzer)

• Breaking: ♻️ Encapsulate the OpenQASM parser in its own library (#822) (@burgholzer)

• Breaking: ♻️ Replace Config template from DD package with constructor argument (#886) (@burgholzer)

• Breaking: ♻️ Remove template parameters from MemoryManager and adjacent classes (#866) (@rotmanjanez)

• Breaking: ♻️ Refactor algorithms to use factory functions instead of inheritance (@a9b7e70) (@burgholzer)

• Breaking: ♻️ Change pointer parameters to references in DD package (#798) (@burgholzer)

• Breaking: ♻️ Change registers from typedef to actual type (#807) (@burgholzer)

• Breaking: ♻️ Refactor NAComputation class hierarchy (#846, #877) (@ystade)

• Breaking: ⬆️ Bump minimum required CMake version to 3.24.0 (#879) (@burgholzer)

• Breaking: ⬆️ Bump minimum required uv version to 0.5.20 (#802) (@burgholzer)

• 📝 Rework existing project documentation (#789, #842) (@burgholzer)

• 📄 Use PEP 639 license expressions (#847) (@burgholzer)

Removed

• Breaking: 🔥 Remove the Teleportation gate from the IR (#882) (@burgholzer)

• Breaking: 🔥 Remove parsers for .real, .qc, .tfc, and GRCS files (#822) (@burgholzer)

• Breaking: 🔥 Remove tensor dump functionality (#798) (@burgholzer)

• Breaking: 🔥 Remove extract_probability_vector functionality (#883) (@burgholzer)

Fixed

• 🐛 Fix Qiskit layout import and handling (#849, #858) (@burgholzer)

• 🐛 Properly handle timing literals in QASM parser (#724) (@burgholzer)

• 🐛 Fix stripping of idle qubits (#763) (@burgholzer)

• 🐛 Fix permutation handling in OpenQASM dump (#810) (@burgholzer)

• 🐛 Fix out-of-bounds error in ZX EdgeIterator (#758) (@burgholzer)

• 🐛 Fix endianness in DCX and XX_minus_YY gate matrix definition (#741) (@burgholzer)

• 🐛 Fix needless dummy register in empty circuit construction (#758) (@burgholzer)

2.7.0 - 2024-10-08

📚 Refer to the GitHub Release Notes for previous changelogs.