Prepare for 0.2.1 release (#59)

* Clarify that Jet is also a Python library

* Improve CudaScopedDevice comments

* Add changelog entry and set development release to 0.2.1

* Save changes to README

* Move PR #49 changelog entry to 0.2.1 release

* Add citations to README and Sphinx documentation

* Add changelog entry for citations

.github/CHANGELOG.md

@@ -1,7 +1,19 @@
-## Release 0.3.0 (development release)
+## Release 0.2.1 (development release)
+
+### New features since last release
+
+* The Jet interpreter for XIR scripts now supports an `expval` output. [(#49)](https://github.com/XanaduAI/jet/pull/49)
+
+### Improvements
+
+* The `CudaTensor` class no longer includes [Taskflow](https://taskflow.github.io/) headers. [(#56)](https://github.com/XanaduAI/jet/pull/56)
 
 ### Documentation
 
+* Links to the [Jet paper](https://arxiv.org/abs/2107.09793) are now included in the README and Sphinx documentation. [(#59)](https://github.com/XanaduAI/jet/pull/59)
+
+* The `CudaScopedDevice` documentation is now consistent with other classes in Jet. [(#59)](https://github.com/XanaduAI/jet/pull/59)
+
 * The documentation section of the development guide now lists system dependencies. [(#52)](https://github.com/XanaduAI/jet/pull/52)
 
 * Static versions of jQuery and Bootstrap are no longer included in the CSS theme. [(#52)](https://github.com/XanaduAI/jet/pull/52)
@@ -12,14 +24,12 @@
 
 This release contains contributions from (in alphabetical order):
 
-[Mikhail Andrenkov](https://github.com/Mandrenkov).
+[Mikhail Andrenkov](https://github.com/Mandrenkov), [Trevor Vincent](https://github.com/trevor-vincent).
 
 ## Release 0.2.0 (current release)
 
 ### New features since last release
 
-* The Jet interpreter for XIR scripts now supports an `expval` output. [(#49)](https://github.com/XanaduAI/jet/pull/49)
-
 * The Jet interpreter for XIR scripts now accepts a `dimension` option for CV circuits. [(#47)](https://github.com/XanaduAI/jet/pull/47)
 
 * The Jet interpreter for XIR programs now supports a `probabilities` output. [(#44)](https://github.com/XanaduAI/jet/pull/44)

README.rst

@@ -22,8 +22,8 @@
     :alt: License
     :target: https://www.apache.org/licenses/LICENSE-2.0
 
-`Jet <https://quantum-jet.readthedocs.io>`_ is a cross-platform C++ library for
-simulating quantum circuits using tensor network contractions.
+`Jet <https://quantum-jet.readthedocs.io>`_ is a cross-platform C++ and Python
+library for simulating quantum circuits using tensor network contractions.
 
 Features
 ========
@@ -173,6 +173,15 @@ Support
 If you are having issues, please let us know by posting the issue on our GitHub
 issue tracker.
 
+Authors
+=======
+
+Jet is the work of `many contributors <https://github.com/XanaduAI/jet/graphs/contributors>`_.
+
+If you are doing research using Jet, please cite our paper:
+
+    Trevor Vincent, Lee J. O'Riordan, Mikhail Andrenkov, Jack Brown, Nathan Killoran, Haoyu Qi, and Ish Dhand. *Jet: Fast quantum circuit simulations with parallel task-based tensor-network contraction.* 2021. `arxiv:2107.09793 <https://arxiv.org/abs/2107.09793>`_
+
 License
 =======
 

docs/dev/research.rst

@@ -4,6 +4,10 @@ Research and contribution
 Research
 --------
 
+If you are doing research using Jet, please cite our paper:
+
+    Trevor Vincent, Lee J. O'Riordan, Mikhail Andrenkov, Jack Brown, Nathan Killoran, Haoyu Qi, and Ish Dhand. *Jet: Fast quantum circuit simulations with parallel task-based tensor-network contraction.* 2021. `arxiv:2107.09793 <https://arxiv.org/abs/2107.09793>`_
+
 We are always open for collaboration, and can be contacted at [email protected].
 
 Contribution

docs/index.rst

@@ -96,6 +96,13 @@ Features
 
 * *Qudits.* Models quantum systems with an **arbitrary number of basis states**.
 
+How to cite
+===========
+
+If you are doing research using Jet, please cite our paper:
+
+    Trevor Vincent, Lee J. O'Riordan, Mikhail Andrenkov, Jack Brown, Nathan Killoran, Haoyu Qi, and Ish Dhand. *Jet: Fast quantum circuit simulations with parallel task-based tensor-network contraction.* 2021. `arxiv:2107.09793 <https://arxiv.org/abs/2107.09793>`_
+
 Support
 =======
 

include/jet/CudaTensorHelpers.hpp

@@ -176,50 +176,57 @@ ReverseVector(const std::vector<DataType> &input)
     return std::vector<DataType>{input.rbegin(), input.rend()};
 }
 
-/** @class CudaScopedDevice
-
-@brief RAII-styled device context switch. Code taken from Taskflow.
-
-%cudaScopedDevice is neither movable nor copyable.
-*/
+/** @brief `%CudaScopedDevice` uses RAII to select a CUDA device context.
+ *
+ * @see https://taskflow.github.io/taskflow/classtf_1_1cudaScopedDevice.html
+ *
+ * @note A `%CudaScopedDevice` instance cannot be moved or copied.
+ *
+ * @warning This class is not thread-safe.
+ */
 class CudaScopedDevice {
 
   public:
     /**
-    @brief constructs a RAII-styled device switcher
-
-    @param device device context to scope in the guard
-    */
+     * @brief Constructs a `%CudaScopedDevice` using a CUDA device.
+     *
+     *  @param device CUDA device to scope in the guard.
+     */
     explicit CudaScopedDevice(int device);
 
     /**
-    @brief destructs the guard and switches back to the previous device context
-    */
+     * @brief Destructs a `%CudaScopedDevice`, switching back to the previous
+     *        CUDA device context.
+     */
     ~CudaScopedDevice();
 
   private:
     CudaScopedDevice() = delete;
     CudaScopedDevice(const CudaScopedDevice &) = delete;
     CudaScopedDevice(CudaScopedDevice &&) = delete;
 
-    int _p;
+    /// The previous CUDA device (or -1 if the device passed to the constructor
+    /// is the current CUDA device).
+    int prev_device_;
 };
 
-inline CudaScopedDevice::CudaScopedDevice(int dev)
+inline CudaScopedDevice::CudaScopedDevice(int device)
 {
-    JET_CUDA_IS_SUCCESS(cudaGetDevice(&_p));
-    if (_p == dev) {
-        _p = -1;
+    JET_CUDA_IS_SUCCESS(cudaGetDevice(&prev_device_));
+    if (prev_device_ == device) {
+        prev_device_ = -1;
     }
     else {
-        JET_CUDA_IS_SUCCESS(cudaSetDevice(dev));
+        JET_CUDA_IS_SUCCESS(cudaSetDevice(device));
     }
 }
 
 inline CudaScopedDevice::~CudaScopedDevice()
 {
-    if (_p != -1) {
-        cudaSetDevice(_p);
+    if (prev_device_ != -1) {
+        // Throwing exceptions from a destructor can be dangerous.
+        // See https://isocpp.org/wiki/faq/exceptions#ctor-exceptions.
+        cudaSetDevice(prev_device_);
     }
 }