Merge branch 'gpuNewAPI_backend' into gpuNewAPI_fullSup

.github/CHANGELOG.md

@@ -2,8 +2,12 @@
 
 ### New features since last release
 
+* Add shot measurement support to `lightning.tensor`.
+  [(#852)](https://github.com/PennyLaneAI/pennylane-lightning/pull/852)
+
 * Build and upload Lightning-Tensor wheels (x86_64, AARCH64) to PyPI.
   [(#862)](https://github.com/PennyLaneAI/pennylane-lightning/pull/862)
+  [(#905)](https://github.com/PennyLaneAI/pennylane-lightning/pull/905)
 
 * Add `Projector` observable support via diagonalization to Lightning-GPU.
   [(#894)](https://github.com/PennyLaneAI/pennylane-lightning/pull/894)
@@ -27,34 +31,46 @@
 
 ### Improvements
 
+* Skip the compilation of Lightning simulators and development requirements to boost the build of public docs up to 5x.
+  [(#904)](https://github.com/PennyLaneAI/pennylane-lightning/pull/904)
+
+* Build Lightning wheels in `Release` mode.
+  [(#903)](https://github.com/PennyLaneAI/pennylane-lightning/pull/903)
+
 * Update Pybind11 to 2.13.5.
   [(#901)](https://github.com/PennyLaneAI/pennylane-lightning/pull/901)
 
 * Migrate wheels artifacts to v4.
   [(#893)](https://github.com/PennyLaneAI/pennylane-lightning/pull/893)
 
-* Prefer `tomlkit` over `toml` for building Lightning wheels, and choose `tomli` and `tomllib` over `toml` when installing the package.
-  [(#857)](https://github.com/PennyLaneAI/pennylane-lightning/pull/857)
-
 * Update GitHub actions in response to a high-severity vulnerability.
   [(#887)](https://github.com/PennyLaneAI/pennylane-lightning/pull/887)
 
+* Optimize and simplify controlled kernels in Lightning-Qubit.
+  [(#882)](https://github.com/PennyLaneAI/pennylane-lightning/pull/882)
+
 * Optimize gate cache recording for `lightning.tensor` C++ layer.
   [(#879)](https://github.com/PennyLaneAI/pennylane-lightning/pull/879)
 
+* Unify Lightning-Kokkos device and Lightning-Qubit device under a Lightning Base device.
+  [(#876)](https://github.com/PennyLaneAI/pennylane-lightning/pull/876)
+
 * Smarter defaults for the `split_obs` argument in the serializer. The serializer splits linear combinations into chunks instead of all their terms.
   [(#873)](https://github.com/PennyLaneAI/pennylane-lightning/pull/873/)
 
-* Unify Lightning-Kokkos device and Lightning-Qubit device under a Lightning Base device.
-  [(#876)](https://github.com/PennyLaneAI/pennylane-lightning/pull/876)
-  
+* Prefer `tomlkit` over `toml` for building Lightning wheels, and choose `tomli` and `tomllib` over `toml` when installing the package.
+  [(#857)](https://github.com/PennyLaneAI/pennylane-lightning/pull/857)
+
 * LightningKokkos gains native support for the `PauliRot` gate.
   [(#855)](https://github.com/PennyLaneAI/pennylane-lightning/pull/855)
 
 ### Documentation
 
 ### Bug fixes
 
+* Bug fix for analytic `probs` in the `lightning.tensor` C++ layer.
+  [(#906)](https://github.com/PennyLaneAI/pennylane-lightning/pull/906)
+
 ### Contributors
 
 This release contains contributions from (in alphabetical order):

.github/workflows/docker_linux_x86_64.yml

@@ -47,7 +47,7 @@ jobs:
     timeout-minutes: 180
     name: docker::${{ matrix.os }}::${{ matrix.pl_backend }}::${{ inputs.lightning-version }}
     runs-on:
-      group: 'Lightning Additional Runners'
+      group: 'PL Additional Runners'
     steps:
 
       - name: Checkout

.github/workflows/wheel_linux_aarch64.yml

@@ -145,7 +145,8 @@ jobs:
             cat /etc/yum.conf | sed "s/\[main\]/\[main\]\ntimeout=5/g" > /etc/yum.conf
             python -m pip install ninja cmake~=3.27.0
 
-          CIBW_ENVIRONMENT: CMAKE_ARGS="-DENABLE_LAPACK=OFF"
+          CIBW_ENVIRONMENT: |
+            CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
 
           CIBW_MANYLINUX_AARCH64_IMAGE: manylinux_2_28
 

.github/workflows/wheel_linux_aarch64_cuda.yml

@@ -79,7 +79,7 @@ jobs:
             PATH=/opt/rh/gcc-toolset-12/root/usr/bin:$PATH:/usr/local/cuda-${{ matrix.cuda_version }}/bin \
             LD_LIBRARY_PATH=/opt/rh/gcc-toolset-12/root/usr/lib64:/opt/rh/gcc-toolset-12/root/usr/lib:/opt/rh/gcc-toolset-12/root/usr/lib64/dyninst:/opt/rh/gcc-toolset-12/root/usr/lib/dyninst:$LD_LIBRARY_PATH:/usr/local/cuda-${{ matrix.cuda_version }}/lib64 \
             PKG_CONFIG_PATH=/opt/rh/gcc-toolset-12/root/usr/lib64/pkgconfig:$PKG_CONFIG_PATH \
-            CMAKE_ARGS="-DENABLE_LAPACK=OFF"
+            CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
 
           CIBW_REPAIR_WHEEL_COMMAND_LINUX: "./bin/auditwheel repair -w {dest_dir} {wheel}"
 
@@ -120,6 +120,8 @@ jobs:
         cuda_version: ["12"]
         cibw_build: ${{ fromJson(needs.set_wheel_build_matrix.outputs.python_version) }}
     runs-on: ubuntu-latest
+    permissions:
+      id-token: write
     if: |
       github.event_name == 'release' ||
       github.ref == 'refs/heads/master'
@@ -130,9 +132,7 @@ jobs:
           name:  ${{ runner.os }}-wheels-${{ matrix.pl_backend }}-${{ fromJson('{ "cp310-*":"py310","cp311-*":"py311","cp312-*":"py312" }')[matrix.cibw_build] }}-${{ matrix.arch }}-cu${{ matrix.cuda_version }}.zip
           path: dist
 
-      - name: Upload wheels to PyPI
+      - name: Upload wheels to TestPyPI
         uses: pypa/gh-action-pypi-publish@release/v1
         with:
-          user: __token__
-          password: ${{ secrets.TEST_PYPI_LGPU_TOKEN }}
           repository_url: https://test.pypi.org/legacy/

.github/workflows/wheel_linux_x86_64.yml

@@ -164,7 +164,8 @@ jobs:
             source /opt/rh/gcc-toolset-13/enable -y
             PATH="/opt/rh/gcc-toolset-13/root/usr/bin:$PATH"
 
-          CIBW_ENVIRONMENT: PATH="/opt/rh/gcc-toolset-13/root/usr/bin:$PATH"
+          CIBW_ENVIRONMENT: |
+            PATH="/opt/rh/gcc-toolset-13/root/usr/bin:$PATH" CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
 
           CIBW_BEFORE_TEST: |
             python -m pip install -r requirements-tests.txt

.github/workflows/wheel_linux_x86_64_cuda.yml

@@ -96,7 +96,7 @@ jobs:
             PATH=/opt/rh/gcc-toolset-12/root/usr/bin:$PATH:/usr/local/cuda-${{ matrix.cuda_version }}/bin \
             LD_LIBRARY_PATH=/opt/rh/gcc-toolset-12/root/usr/lib64:/opt/rh/gcc-toolset-12/root/usr/lib:/opt/rh/gcc-toolset-12/root/usr/lib64/dyninst:/opt/rh/gcc-toolset-12/root/usr/lib/dyninst:$LD_LIBRARY_PATH:/usr/local/cuda-${{ matrix.cuda_version }}/lib64 \
             PKG_CONFIG_PATH=/opt/rh/gcc-toolset-12/root/usr/lib64/pkgconfig:$PKG_CONFIG_PATH \
-            CMAKE_ARGS="-DENABLE_LAPACK=OFF"
+            CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
 
           CIBW_REPAIR_WHEEL_COMMAND_LINUX: "./bin/auditwheel repair -w {dest_dir} {wheel}"
 
@@ -138,6 +138,8 @@ jobs:
         cuda_version: ["12"]
         cibw_build: ${{ fromJson(needs.set_wheel_build_matrix.outputs.python_version) }}
     runs-on: ubuntu-latest
+    permissions:
+      id-token: write
     if: |
       github.event_name == 'release' ||
       github.ref == 'refs/heads/master'
@@ -148,9 +150,7 @@ jobs:
           name:  ${{ runner.os }}-wheels-${{ matrix.pl_backend }}-${{ fromJson('{ "cp310-*":"py310","cp311-*":"py311","cp312-*":"py312" }')[matrix.cibw_build] }}-${{ matrix.arch }}-cu${{ matrix.cuda_version }}.zip
           path: dist
 
-      - name: Upload wheels to PyPI
+      - name: Upload wheels to TestPyPI
         uses: pypa/gh-action-pypi-publish@release/v1
         with:
-          user: __token__
-          password: ${{ secrets.TEST_PYPI_LGPU_TOKEN }}
           repository-url: https://test.pypi.org/legacy/

.github/workflows/wheel_macos_arm64.yml

@@ -99,7 +99,7 @@ jobs:
             python -m pip install pybind11 ninja cmake~=3.27.0 setuptools scipy
 
           CIBW_ENVIRONMENT: |
-            CMAKE_ARGS="-DCMAKE_CXX_COMPILER_TARGET=arm64-apple-macos11 -DCMAKE_SYSTEM_NAME=Darwin -DCMAKE_SYSTEM_PROCESSOR=ARM64 -DENABLE_OPENMP=OFF"
+            CMAKE_ARGS="-DCMAKE_CXX_COMPILER_TARGET=arm64-apple-macos11 -DCMAKE_SYSTEM_NAME=Darwin -DCMAKE_SYSTEM_PROCESSOR=ARM64 -DENABLE_OPENMP=OFF -DCMAKE_BUILD_TYPE=Release"
 
           CIBW_BEFORE_TEST: |
             python -m pip install -r requirements-tests.txt

.github/workflows/wheel_macos_x86_64.yml

@@ -147,6 +147,9 @@ jobs:
           CIBW_BEFORE_BUILD: |
             python -m pip install pybind11 ninja cmake~=3.27.0 setuptools scipy
 
+          CIBW_ENVIRONMENT: |
+            CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
+
           PL_BACKEND: ${{ matrix.pl_backend }}
 
           CIBW_BEFORE_TEST: |

.github/workflows/wheel_win_x86_64.yml

@@ -55,7 +55,7 @@ jobs:
         uses: actions/cache@v4
         with:
           path: D:\a\install_dir\${{ matrix.exec_model }}
-          key: ${{ matrix.os }}-kokkos${{ matrix.kokkos_version }}-${{ matrix.exec_model }}-RelWithDebInfo
+          key: ${{ matrix.os }}-kokkos${{ matrix.kokkos_version }}-${{ matrix.exec_model }}-Release
 
       - name: Clone Kokkos libs
         if: steps.kokkos-cache.outputs.cache-hit != 'true'
@@ -80,10 +80,10 @@ jobs:
                           -DKokkos_ENABLE_DEPRECATION_WARNINGS=OFF `
                           -DCMAKE_CXX_STANDARD=20 `
                           -DCMAKE_POSITION_INDEPENDENT_CODE=ON `
-                          -DCMAKE_BUILD_TYPE=RelWithDebInfo `
+                          -DCMAKE_BUILD_TYPE=Release `
                           -T clangcl
-          cmake --build ./Build --config RelWithDebInfo --verbose
-          cmake --install ./Build --config RelWithDebInfo --verbose
+          cmake --build ./Build --config Release --verbose
+          cmake --install ./Build --config Release --verbose
 
   win-wheels:
     needs: [set_wheel_build_matrix, build_dependencies]
@@ -136,7 +136,7 @@ jobs:
             python -m pip install pybind11 cmake~=3.27.0 build
 
           CIBW_ENVIRONMENT: |
-            CMAKE_ARGS="-DENABLE_LAPACK=OFF"
+            CMAKE_ARGS="-DCMAKE_BUILD_TYPE=Release"
 
           CIBW_MANYLINUX_X86_64_IMAGE: manylinux2014
 

.readthedocs.yml

@@ -3,32 +3,17 @@ version: 2
 sphinx:
   configuration: doc/conf.py
 
-python:
-  install:
-    - requirements: ci_build_requirements.txt
-    - requirements: doc/requirements.txt
-    - requirements: requirements-dev.txt
-    - method: pip
-      path: .
-
 build:
   os: ubuntu-22.04
   tools:
     python: "3.10"
   apt_packages:
-    - cmake
-    - build-essential
-    - libopenblas-base
-    - libopenblas-dev
     - graphviz
-    - wget
   jobs:
-    pre_install:
-      - wget https://developer.download.nvidia.com/compute/cuda/12.3.2/local_installers/cuda_12.3.2_545.23.08_linux.run
-      - sh cuda_12.3.2_545.23.08_linux.run --silent --toolkit --toolkitpath=${READTHEDOCS_VIRTUALENV_PATH}/cuda-12.3 || cat /tmp/cuda-installer.log
-      - echo "setuptools~=66.0" >> ci_build_requirements.txt
     post_install:
-      - rm -rf ./build && export PATH=${READTHEDOCS_VIRTUALENV_PATH}/cuda-12.3/bin${PATH:+:${PATH}} && export LD_LIBRARY_PATH=${READTHEDOCS_VIRTUALENV_PATH}/cuda-12.3/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} && PL_BACKEND="lightning_gpu" python scripts/configure_pyproject_toml.py && CMAKE_ARGS="-DPL_DISABLE_CUDA_SAFETY=1" python -m build
-      - rm -rf ./build && PL_BACKEND="lightning_kokkos" python scripts/configure_pyproject_toml.py && python -m build
-      - rm -rf ./build && export PATH=${READTHEDOCS_VIRTUALENV_PATH}/cuda-12.3/bin${PATH:+:${PATH}} && export LD_LIBRARY_PATH=${READTHEDOCS_VIRTUALENV_PATH}/cuda-12.3/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} && PL_BACKEND="lightning_tensor" python scripts/configure_pyproject_toml.py && CMAKE_ARGS="-DPL_DISABLE_CUDA_SAFETY=1" python -m build
+      - python -m pip install --exists-action=w --no-cache-dir -r doc/requirements.txt
+      - PL_BACKEND="lightning_qubit" python scripts/configure_pyproject_toml.py && SKIP_COMPILATION=True python -m build
+      - rm -rf ./build && PL_BACKEND="lightning_gpu" python scripts/configure_pyproject_toml.py && SKIP_COMPILATION=True python -m build
+      - rm -rf ./build && PL_BACKEND="lightning_kokkos" python scripts/configure_pyproject_toml.py && SKIP_COMPILATION=True python -m build
+      - rm -rf ./build && PL_BACKEND="lightning_tensor" python scripts/configure_pyproject_toml.py && SKIP_COMPILATION=True python -m build
       - python -m pip install ./dist/*.whl

Makefile

@@ -63,11 +63,15 @@ clean:
 	rm -rf pennylane_lightning/*_ops*
 	rm -rf *.egg-info
 
-.PHONY: python
+.PHONY: python python-skip-compile
 python:
 	PL_BACKEND=$(PL_BACKEND) python scripts/configure_pyproject_toml.py
 	pip install -e . --config-settings editable_mode=compat -vv
 
+python-skip-compile:
+	PL_BACKEND=$(PL_BACKEND) python scripts/configure_pyproject_toml.py
+	SKIP_COMPILATION=True pip install -e . --config-settings editable_mode=compat -vv
+
 .PHONY: wheel
 wheel:
 	PL_BACKEND=$(PL_BACKEND) python scripts/configure_pyproject_toml.py

doc/conf.py

@@ -21,6 +21,7 @@
 # add these directories to sys.path here. If the directory is relative to the
 # documentation root, use os.path.abspath to make it absolute, like shown here.
 sys.path.insert(0, os.path.abspath(""))
+sys.path.insert(0, os.path.abspath("."))
 sys.path.insert(0, os.path.abspath("_ext"))
 sys.path.insert(0, os.path.join(os.path.dirname(os.path.abspath("doc")), "doc"))
 
@@ -189,6 +190,8 @@ def __getattr__(cls, name):
 # This patterns also effect to html_static_path and html_extra_path
 exclude_patterns = ["_build", "Thumbs.db", ".DS_Store"]
 
+nbsphinx_execute = "never"
+
 # If true, sectionauthor and moduleauthor directives will be shown in the
 # output. They are ignored by default.
 show_authors = True

doc/requirements.txt

@@ -1,3 +1,5 @@
+tomlkit
+build
 breathe
 docutils==0.16
 exhale>=0.3.3

pennylane_lightning/core/src/bindings/Bindings.hpp

@@ -746,7 +746,27 @@ void registerLightningTensorBackendAgnosticMeasurements(PyClass &pyclass) {
             [](MeasurementsT &M, const std::shared_ptr<ObservableT> &ob) {
                 return M.var(*ob);
             },
-            "Variance of an observable object.");
+            "Variance of an observable object.")
+        .def("generate_samples", [](MeasurementsT &M,
+                                    const std::vector<std::size_t> &wires,
+                                    const std::size_t num_shots) {
+            constexpr auto sz = sizeof(std::size_t);
+            const std::size_t num_wires = wires.size();
+            const std::size_t ndim = 2;
+            const std::vector<std::size_t> shape{num_shots, num_wires};
+            auto &&result = M.generate_samples(wires, num_shots);
+
+            const std::vector<std::size_t> strides{sz * num_wires, sz};
+            // return 2-D NumPy array
+            return py::array(py::buffer_info(
+                result.data(), /* data as contiguous array  */
+                sz,            /* size of one scalar        */
+                py::format_descriptor<std::size_t>::format(), /* data type */
+                ndim,   /* number of dimensions      */
+                shape,  /* shape of the matrix       */
+                strides /* strides for each axis     */
+                ));
+        });
 }
 
 /**

pennylane_lightning/core/src/simulators/lightning_qubit/gates/GateIndices.cpp

@@ -47,4 +47,37 @@ auto generateBitPatterns(const std::vector<std::size_t> &qubitIndices,
     }
     return indices;
 }
+
+/**
+ * @brief Introduce quantum controls in indices generated by
+ * generateBitPatterns.
+ *
+ * @param indices Indices for the operation.
+ * @param num_qubits Number of qubits in register.
+ * @param controlled_wires Control wires.
+ * @param controlled_values Control values (false or true).
+ */
+void controlBitPatterns(std::vector<std::size_t> &indices,
+                        const std::size_t num_qubits,
+                        const std::vector<std::size_t> &controlled_wires,
+                        const std::vector<bool> &controlled_values) {
+    constexpr std::size_t one{1U};
+    if (controlled_wires.empty()) {
+        return;
+    }
+    std::vector<std::size_t> controlled_values_i(controlled_values.size());
+    std::transform(controlled_values.begin(), controlled_values.end(),
+                   controlled_values_i.begin(),
+                   [](const bool v) { return static_cast<std::size_t>(v); });
+    std::for_each(
+        indices.begin(), indices.end(),
+        [num_qubits, &controlled_wires, &controlled_values_i](std::size_t &i) {
+            for (std::size_t k = 0; k < controlled_wires.size(); k++) {
+                const std::size_t rev_wire =
+                    (num_qubits - 1) - controlled_wires[k];
+                const std::size_t value = controlled_values_i[k];
+                i = (i & ~(one << rev_wire)) | (value << rev_wire);
+            }
+        });
+}
 } // namespace Pennylane::LightningQubit::Gates

pennylane_lightning/core/src/simulators/lightning_qubit/gates/GateIndices.hpp

@@ -56,6 +56,19 @@ auto getIndicesAfterExclusion(const std::vector<std::size_t> &indicesToExclude,
 auto generateBitPatterns(const std::vector<std::size_t> &qubitIndices,
                          std::size_t num_qubits) -> std::vector<std::size_t>;
 
+/**
+ * @brief Introduce quantum controls in indices generated by
+ * generateBitPatterns.
+ *
+ * @param indices Indices for the operation.
+ * @param num_qubits Number of qubits in register.
+ * @param controlled_wires Control wires.
+ * @param controlled_values Control values (false or true).
+ */
+void controlBitPatterns(std::vector<std::size_t> &indices,
+                        std::size_t num_qubits,
+                        const std::vector<std::size_t> &controlled_wires,
+                        const std::vector<bool> &controlled_values);
 /**
  * @brief Internal utility struct to track data indices of application for
  * operations.