Release Notes
=============

.. raw:: html

    <div class="section" id="version-2-1-0">
    <h2>Version 2.1.0<a class="headerlink" href="#version-2-1-0" title="Permalink to this headline">¶</a></h2>
    <p>This release of Accelerate is a feature release, containing two additions:</p>
    <ul class="simple">
    <li>Profiling tools:<ul>
    <li>An enhanced version of the Python profiler that captures function arguments,
    including shapes and dtypes of Numpy arrays.</li>
    <li>Integration of tools for visualising profiles in Jupyter notebooks, allowing
    interactive experimentation.</li>
    </ul>
    </li>
    <li>Accelerated UFuncs, which provide a speed improvement over Numpy&#8217;s built-in
    UFuncs by using Intel MKL.</li>
    </ul>
    </div>
    <div class="section" id="version-2-0-2">
    <h2>Version 2.0.2<a class="headerlink" href="#version-2-0-2" title="Permalink to this headline">¶</a></h2>
    <p>This update adds support for updated versions of Numba and MKL:</p>
    <ul class="simple">
    <li>Numba 0.23 is now supported, and is a requirement.</li>
    <li>MKL 11.3.1, standard in Anaconda 2.5, is now supported.</li>
    </ul>
    </div>
    <div class="section" id="version-2-0-1">
    <h2>Version 2.0.1<a class="headerlink" href="#version-2-0-1" title="Permalink to this headline">¶</a></h2>
    <p>This minor update to Accelerate adds no functional changes, but provides
    additional clarification of the the relationship between the Accelerate and
    NumbaPro conda packages upon installation.</p>
    </div>
    <div class="section" id="version-2-0">
    <h2>Version 2.0<a class="headerlink" href="#version-2-0" title="Permalink to this headline">¶</a></h2>
    <p>NumbaPro has been deprecated, and its code generation features have been moved
    into open-source Numba. The CUDA library functions have been moved into
    Accelerate, along with some Intel MKL functionality. High-level functions and
    access to additional native library implementations will be added in future
    releases of Accelerate, and there will be no further updates to NumbaPro.</p>
    <p>A NumbaPro compatibility layer (listed as release 0.22.0 of NumbaPro) provides
    access to the new Accelerate packages through the old NumbaPro package names.
    This avoids the need to change any existing code immediately for use with
    Accelerate. A warning will be generated upon import of the compatibility layer,
    to highlight the deprecation of the NumbaPro package.</p>
    <div class="section" id="cuda-libraries">
    <h3>CUDA Libraries<a class="headerlink" href="#cuda-libraries" title="Permalink to this headline">¶</a></h3>
    <p>CUDA library functionality is equivalent to that in NumbaPro 0.21, with the
    following packages renamed:</p>
    <table border="1" class="docutils">
    <colgroup>
    <col width="52%" />
    <col width="48%" />
    </colgroup>
    <thead valign="bottom">
    <tr class="row-odd"><th class="head">NumbaPro Package</th>
    <th class="head">Accelerate package</th>
    </tr>
    </thead>
    <tbody valign="top">
    <tr class="row-even"><td><code class="docutils literal"><span class="pre">numbapro.cudalib.cublas</span></code></td>
    <td><code class="docutils literal"><span class="pre">accelerate.cuda.blas</span></code></td>
    </tr>
    <tr class="row-odd"><td><code class="docutils literal"><span class="pre">numbapro.cudalib.cufft</span></code></td>
    <td><code class="docutils literal"><span class="pre">accelerate.cuda.fft</span></code></td>
    </tr>
    <tr class="row-even"><td><code class="docutils literal"><span class="pre">numbapro.cudalib.curand</span></code></td>
    <td><code class="docutils literal"><span class="pre">accelerate.cuda.rand</span></code></td>
    </tr>
    <tr class="row-odd"><td><code class="docutils literal"><span class="pre">numbapro.cudalib.cusparse</span></code></td>
    <td><code class="docutils literal"><span class="pre">accelerate.cuda.sparse</span></code></td>
    </tr>
    <tr class="row-even"><td><code class="docutils literal"><span class="pre">numbapro.cudalib.sorting</span></code></td>
    <td><code class="docutils literal"><span class="pre">accelerate.cuda.sorting</span></code></td>
    </tr>
    </tbody>
    </table>
    </div>
    <div class="section" id="code-generation">
    <h3>Code Generation<a class="headerlink" href="#code-generation" title="Permalink to this headline">¶</a></h3>
    <p>The <code class="docutils literal"><span class="pre">vectorize</span></code> targets <code class="docutils literal"><span class="pre">parallel</span></code> and <code class="docutils literal"><span class="pre">cuda</span></code> can now be accessed
    with Numba, as can the <code class="docutils literal"><span class="pre">cuda.reduce</span></code> decorator. Printing of integers and
    floating point values from CUDA kernels is also possible in Numba, and no longer
    requires NumbaPro or Accelerate to be imported.</p>
    </div>
    <div class="section" id="intel-mkl">
    <h3>Intel MKL<a class="headerlink" href="#intel-mkl" title="Permalink to this headline">¶</a></h3>
    <p>The MKL FFT implementation is available in the package
    <code class="docutils literal"><span class="pre">accelerate.mkl.fftpack</span></code>, and MKL service functions are available in the
    <code class="docutils literal"><span class="pre">accelerate.mkl</span></code> package.</p>
    </div>
