Merge branch 'release/v2.0.0'.

.github/workflows/python-publish.yml

@@ -1,31 +1,40 @@
 # This workflow will upload a Python Package using Twine when a release is created
-# For more information see: https://help.github.com/en/actions/language-and-framework-guides/using-python-with-github-actions#publishing-to-package-registries
+# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python#publishing-to-package-registries
+
+# This workflow uses actions that are not certified by GitHub.
+# They are provided by a third-party and are governed by
+# separate terms of service, privacy policy, and support
+# documentation.
 
 name: Upload Python Package
 
 on:
   release:
-    types: [created]
+    types: [published]
+  workflow_dispatch:
+
+permissions:
+  contents: read
 
 jobs:
   deploy:
 
     runs-on: ubuntu-latest
 
     steps:
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v3
     - name: Set up Python
-      uses: actions/setup-python@v2
+      uses: actions/setup-python@v3
       with:
         python-version: '3.x'
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip
-        pip install setuptools wheel twine
-    - name: Build and publish
-      env:
-        TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
-        TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
-      run: |
-        python setup.py sdist bdist_wheel
-        twine upload dist/*
+        pip install build
+    - name: Build package
+      run: python -m build
+    - name: Publish package
+      uses: pypa/gh-action-pypi-publish@27b31702a0e7fc50959f5ad993c78deac1bdfc29
+      with:
+        user: __token__
+        password: ${{ secrets.PYPI_API_TOKEN }}

AUTHORS

@@ -7,7 +7,7 @@ Arvind Iyer <a.iyer@imperial.ac.uk>
 Niki Loppi <n.loppi15@imperial.ac.uk>
 George Ntemos <george.ntemos11@imperial.ac.uk>
 Jin Seok Park <jin-seok.park@imperial.ac.uk>
-Will Trojak <wt247@tamu.edu>
+Will Trojak <w.trojak@ibm.com>
 Brian Vermeire <b.vermeire@imperial.ac.uk>
 Peter Vincent <p.vincent@imperial.ac.uk>
 Freddie Witherden <freddie@witherden.org>

doc/src/backends/backend-cuda.rst

@@ -0,0 +1,38 @@
+**************
+[backend-cuda]
+**************
+
+Parameterises the CUDA backend with
+
+1. ``device-id`` --- method for selecting which device(s) to run on:
+
+     *int* | ``round-robin`` | ``local-rank`` | ``uuid``
+
+2. ``mpi-type`` --- type of MPI library that is being used:
+
+     ``standard`` | ``cuda-aware``
+
+3. ``cflags`` --- additional NVIDIA realtime compiler (``nvrtc``) flags:
+
+    *string*
+
+4. ``cublas-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 512:
+
+    *int*
+
+5. ``gimmik-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 8:
+
+    *int*
+
+6. ``gimmik-nbench`` --- number of benchmarking runs for each
+   kernel, defaults to 5:
+
+    *int*
+
+Example::
+
+    [backend-cuda]
+    device-id = round-robin
+    mpi-type = standard

doc/src/backends/backend-hip.rst

@@ -0,0 +1,34 @@
+*************
+[backend-hip]
+*************
+
+Parameterises the HIP backend with
+
+1. ``device-id`` --- method for selecting which device(s) to run on:
+
+     *int* | ``local-rank`` | ``uuid``
+
+2. ``mpi-type`` --- type of MPI library that is being used:
+
+     ``standard`` | ``hip-aware``
+
+3. ``rocblas-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 2048:
+
+     *int*
+
+4. ``gimmik-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 8:
+
+    *int*
+
+5. ``gimmik-nbench`` --- number of benchmarking runs for each
+   kernel, defaults to 5:
+
+     *int*
+
+Example::
+
+    [backend-hip]
+    device-id = local-rank
+    mpi-type = standard

doc/src/backends/backend-metal.rst

@@ -0,0 +1,25 @@
+****************
+[backend-metal]
+****************
+
+Parameterises the Metal backend with
+
+1. ``gimmik-max-nnz`` --- cutoff for GiMMiK in terms of the number of
+   non-zero entires in a constant matrix, defaults to 2048:
+
+    *int*
+
+2. ``gimmik-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 18:
+
+    *int*
+
+3. ``gimmik-nbench`` --- number of benchmarking runs for each
+   kernel, defaults to 40:
+
+     *int*
+
+Example::
+
+    [backend-metal]
+    gimmik-max-nnz = 512

doc/src/backends/backend-opencl.rst

@@ -0,0 +1,40 @@
+****************
+[backend-opencl]
+****************
+
+Parameterises the OpenCL backend with
+
+1. ``platform-id`` --- for selecting platform id:
+
+    *int* | *string*
+
+2. ``device-type`` --- for selecting what type of device(s) to run on:
+
+    ``all`` | ``cpu`` | ``gpu`` | ``accelerator``
+
+3. ``device-id`` --- for selecting which device(s) to run on:
+
+    *int* | *string* | ``local-rank`` | ``uuid``
+
+4. ``gimmik-max-nnz`` --- cutoff for GiMMiK in terms of the number of
+   non-zero entires in a constant matrix, defaults to 2048:
+
+    *int*
+
+5. ``gimmik-nkerns`` --- number of kernel algorithms to try when
+   benchmarking, defaults to 8:
+
+    *int*
+
+6. ``gimmik-nbench`` --- number of benchmarking runs for each
+   kernel, defaults to 5:
+
+     *int*
+
+Example::
+
+    [backend-opencl]
+    platform-id = 0
+    device-type = gpu
+    device-id = local-rank
+    gimmik-max-nnz = 512

doc/src/backends/backend-openmp.rst

@@ -0,0 +1,29 @@
+****************
+[backend-openmp]
+****************
+
+Parameterises the OpenMP backend with
+
+1. ``cc`` --- C compiler:
+
+    *string*
+
+2. ``cflags`` --- additional C compiler flags:
+
+    *string*
+
+3. ``alignb`` --- alignment requirement in bytes; must be a power of
+   two and at least 32:
+
+    *int*
+
+4. ``schedule`` --- OpenMP loop scheduling scheme:
+
+    ``static`` | ``dynamic`` | ``dynamic, n`` | ``guided`` | ``guided, n``
+
+    where *n* is a positive integer.
+
+Example::
+
+    [backend-openmp]
+    cc = gcc

doc/src/backends/backend.rst

@@ -0,0 +1,33 @@
+*********
+[backend]
+*********
+
+Parameterises the backend with
+
+1. ``precision`` --- number precision, note not all backends support 
+   double precision:
+
+    ``single`` | ``double``
+
+2. ``memory-model`` --- if to enable support for large working sets;
+    should be ``normal`` unless a memory-model error is encountered:
+
+    ``normal`` | ``large``
+
+3. ``rank-allocator`` --- MPI rank allocator:
+
+    ``linear`` | ``random``
+
+4. ``collect-wait-times`` --- if to track MPI request wait times or not:
+
+    ``True`` | ``False``
+
+5. ``collect-wait-times-len`` --- size of the wait time history buffer:
+
+     *int*
+
+Example::
+
+    [backend]
+    precision = double
+    rank-allocator = linear