.github/workflows/ci.yml

+name: CI
+on:
+    push:
+        branches:
+        - master
+    pull_request:
+        paths-ignore:
+        - 'doc/*.rst'
+    schedule:
+        - cron:  '17 3 * * 0'
+
+jobs:
+    flake8:
+        name: Flake8
+        runs-on: ubuntu-latest
+        steps:
+        -   uses: actions/checkout@v2
+        -
+            uses: actions/setup-python@v1
+            with:
+                python-version: '3.x' 
+        -   name: "Main Script"
+            run: |
+                curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/prepare-and-run-flake8.sh
+                . ./prepare-and-run-flake8.sh ./loopy ./test
+
+    pylint:
+        name: Pylint
+        runs-on: ubuntu-latest
+        steps:
+        -   uses: actions/checkout@v2
+        -   name: "Main Script"
+            run: |
+                sed 's/python=3/python=3.7/' .test-conda-env-py3.yml > .test-conda-env.yml
+                CONDA_ENVIRONMENT=.test-conda-env.yml
+                USE_CONDA_BUILD=1
+                curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/prepare-and-run-pylint.sh
+                . ./prepare-and-run-pylint.sh loopy test/test_*.py
+
+    pytest3:
+        name: Conda Pytest Py3
+        runs-on: ubuntu-latest
+        steps:
+        -   uses: actions/checkout@v2
+        -   name: "Main Script"
+            run: |
+                CONDA_ENVIRONMENT=.test-conda-env-py3.yml
+                curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project-within-miniconda.sh
+                . ./build-and-test-py-project-within-miniconda.sh
+
+    pytest_twice:
+        name: Pytest twice (for cache behavior) on Py${{ matrix.python-version }}
+        runs-on: ubuntu-latest
+        steps:
+        -   uses: actions/checkout@v2
+        -   name: "Main Script"
+            run: |
+                CONDA_ENVIRONMENT=.test-conda-env-py3.yml
+                curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project-within-miniconda.sh
+                . ./build-and-test-py-project-within-miniconda.sh
+                ${PY_EXE} -m pytest -rw --durations=10 --tb=native  --junitxml=pytest.xml -rxs $TESTABLES
+
+# vim: sw=4

.gitlab-ci.yml

-Python 2.7 POCL:
-  script:
-  - export PY_EXE=python2.7
-  - export PYOPENCL_TEST=portable
-  - export EXTRA_INSTALL="pybind11 numpy mako"
-  - export LOOPY_NO_CACHE=1
-  - export NO_DOCTESTS=1
-  - curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project.sh
-  - ". ./build-and-test-py-project.sh"
-  tags:
-  - python2.7
-  - pocl
-  except:
-  - tags
-  artifacts:
-    reports:
-      junit: test/pytest.xml
-
-
 Python 3 POCL:
   script:
   - export PY_EXE=python3
-  - export PYOPENCL_TEST=portable
+  - export PYOPENCL_TEST=portable:pthread
   - export EXTRA_INSTALL="pybind11 numpy mako"
   - export LOOPY_NO_CACHE=1
   - curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project.sh
@@ -57,7 +38,7 @@ Python 3 Intel:
 Python 3 POCL Twice With Cache:
   script:
   - export PY_EXE=python3
-  - export PYOPENCL_TEST=portable
+  - export PYOPENCL_TEST=portable:pthread
   - export EXTRA_INSTALL="pybind11 numpy mako"
   - curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project.sh
   - ". ./build-and-test-py-project.sh"
@@ -77,7 +58,7 @@ Python 3 POCL Twice With Cache:
 # PyPy POCL:
 #   script:
 #   - export PY_EXE=pypy
-#   - export PYOPENCL_TEST=portable
+#   - export PYOPENCL_TEST=portable:pthread
 #   - export EXTRA_INSTALL="pybind11 numpy mako"
 #   - curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project.sh
 #   - ". ./build-and-test-py-project.sh"
@@ -90,7 +71,7 @@ Python 3 POCL Twice With Cache:
 Python 3 POCL Examples:
   script:
   - export PY_EXE=python3
-  - export PYOPENCL_TEST=portable
+  - export PYOPENCL_TEST=portable:pthread
   - export EXTRA_INSTALL="pybind11 numpy mako pyvisfile matplotlib ipykernel nbconvert"
   - ". ./build-py-project-and-run-examples.sh"
   tags:
@@ -114,20 +95,6 @@ Pylint:
   except:
   - tags
 
-CentOS binary:
-  script:
-  - (cd build-helpers; ./make-linux-build-docker.sh --nodate)
-  - (cd ./build-helpers; ./loopy-centos6 ../examples/fortran/sparse.floopy)
-  artifacts:
-    expire_in: 4 weeks
-    paths:
-    - build-helpers/loopy-centos6
-  tags:
-  - docker
-  only:
-  - master
-  retry: 2
-
 Documentation:
   script:
   - EXTRA_INSTALL="pybind11 numpy"

.test-conda-env-py3.yml

 name: test-conda-env
 channels:
 - conda-forge
-- defaults
+- nodefaults
 
 dependencies:
 - python=3
 - git
-- conda-forge::numpy
+- numpy
 - pocl
 - mako
 - pyopencl

README.rst

@@ -4,9 +4,9 @@ Loopy: Transformation-Based Generation of High-Performance CPU/GPU Code
 .. image:: https://gitlab.tiker.net/inducer/loopy/badges/master/pipeline.svg
     :alt: Gitlab Build Status
     :target: https://gitlab.tiker.net/inducer/loopy/commits/master
-.. image:: https://dev.azure.com/ak-spam/inducer/_apis/build/status/inducer.loopy?branchName=master
-    :alt: Azure Build Status
-    :target: https://dev.azure.com/ak-spam/inducer/_build/latest?definitionId=10&branchName=master
+.. image:: https://github.com/inducer/loopy/workflows/CI/badge.svg?branch=master&event=push
+    :alt: Github Build Status
+    :target: https://github.com/inducer/loopy/actions?query=branch%3Amaster+workflow%3ACI+event%3Apush
 .. image:: https://badge.fury.io/py/loo.py.png
     :alt: Python Package Index Release Page
     :target: https://pypi.org/project/loo.py/

azure-pipelines.yml

-jobs:
--
-    job: 'Python2'
-    pool:
-        vmImage: 'ubuntu-latest'
-
-    steps:
-    -
-        script: |
-            set -e
-            sed 's/python=3/python=2.7/' .test-conda-env-py3.yml > .test-conda-env-py2.yml
-            cat .test-conda-env-py2.yml
-            CONDA_ENVIRONMENT=.test-conda-env-py2.yml
-            curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project-within-miniconda.sh
-            . ./build-and-test-py-project-within-miniconda.sh
-
-        displayName: 'Pytest Conda'
-    -
-        task: PublishTestResults@2
-        inputs:
-            testResultsFormat: 'JUnit'
-            testResultsFiles: 'test/pytest.xml'
-
--
-    job: 'Python3'
-    pool:
-        vmImage: 'ubuntu-latest'
-
-    steps:
-    -
-        script: |
-            set -e
-            CONDA_ENVIRONMENT=.test-conda-env-py3.yml
-            curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project-within-miniconda.sh
-            . ./build-and-test-py-project-within-miniconda.sh
-
-        displayName: 'Pytest Conda'
-
-    -
-        task: PublishTestResults@2
-        inputs:
-            testResultsFormat: 'JUnit'
-            testResultsFiles: 'test/pytest.xml'
-
--
-    job: 'Python3Twice'
-    displayName: "Python3 - run tests twice to test cache behavior"
-    pool:
-        vmImage: 'ubuntu-latest'
-
-    steps:
-    -
-        script: |
-            set -e
-            CONDA_ENVIRONMENT=.test-conda-env-py3.yml
-            curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/build-and-test-py-project-within-miniconda.sh
-            . ./build-and-test-py-project-within-miniconda.sh
-            ${PY_EXE} -m pytest -rw --durations=10 --tb=native  --junitxml=pytest.xml -rxs $TESTABLES
-
-        displayName: 'Pytest Conda'
-
-    -
-        task: PublishTestResults@2
-        inputs:
-            testResultsFormat: 'JUnit'
-            testResultsFiles: 'test/pytest.xml'
-
--
-    job: 'Flake8'
-    pool:
-        vmImage: 'ubuntu-latest'
-    strategy:
-        matrix:
-            Python37:
-                python.version: '3.7'
-
-    steps:
-    -
-        task: UsePythonVersion@0
-        inputs:
-            versionSpec: '$(python.version)'
-
-    -
-        script: |
-            set -e
-            curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/prepare-and-run-flake8.sh
-            . ./prepare-and-run-flake8.sh loopy test
-
-        displayName: 'Flake8'
-
--
-    job: 'Pylint'
-    pool:
-        vmImage: 'ubuntu-latest'
-
-    steps:
-    -
-        script: |
-            set -e
-            sed 's/python=3/python=3.7/' .test-conda-env-py3.yml > .test-conda-env.yml
-            CONDA_ENVIRONMENT=.test-conda-env.yml
-            USE_CONDA_BUILD=1
-            curl -L -O -k https://gitlab.tiker.net/inducer/ci-support/raw/master/prepare-and-run-pylint.sh
-            . ./prepare-and-run-pylint.sh loopy test/test_*.py
-
-        displayName: 'Pylint'
-
-schedules:
--
-    cron: "0 0 * * 0"
-    displayName: Weekly build
-    branches:
-        include:
-        - master

build-helpers/.gitignore

-loopy-*-20[0-9][0-9]*

build-helpers/loopy.spec

-# -*- mode: python -*-
-
-from os.path import basename, dirname, join
-from glob import glob
-
-single_file = True
-
-# This makes the executable spew debug info.
-debug = False
-
-from os.path import expanduser
-
-import packaging # pip install packaging to add
-
-a = Analysis(['../bin/loopy'],
-             pathex=[expanduser('~/src/loopy')],
-             hiddenimports=[
-                "decorator",
-                "appdirs",
-                "packaging.markers",
-                "packaging.specifiers",
-                "packaging.version",
-                "packaging.requirements",
-                ],
-             hookspath=None,
-             runtime_hooks=None,
-             excludes=["hedge", "meshpy", "pyopencl", "PIL"]
-             )
-
-import ply.lex
-import ply.yacc
-
-
-a.datas += [
-  (join("py-src", "ply", "lex", basename(fn)), fn, "DATA")
-  for fn in glob(join(dirname(ply.lex.__file__), "*.py"))
-  ] + [
-  (join("py-src", "ply", "yacc", basename(fn)), fn, "DATA")
-  for fn in glob(join(dirname(ply.yacc.__file__), "*.py"))
-  ]
-
-pyz = PYZ(a.pure)
-
-if single_file:
-    exe = EXE(pyz,
-              a.scripts,
-              a.binaries,
-              a.zipfiles,
-              a.datas,
-              name='loopy',
-              debug=debug,
-              strip=None,
-              upx=True,
-              console=True)
-else:
-    exe = EXE(pyz,
-              a.scripts,
-              exclude_binaries=True,
-              name='loopy',
-              debug=debug,
-              strip=None,
-              upx=True,
-              console=True)
-    coll = COLLECT(exe,
-                   a.binaries,
-                   a.zipfiles,
-                   a.datas,
-                   strip=None,
-                   upx=True,
-                   name='loopy')

build-helpers/make-linux-build-docker-inner-part-2.sh

-#! /bin/bash
-
-set -e
-set -x
-
-VENV_VERSION="virtualenv-15.2.0"
-rm -Rf "$VENV_VERSION"
-curl -k https://files.pythonhosted.org/packages/b1/72/2d70c5a1de409ceb3a27ff2ec007ecdd5cc52239e7c74990e32af57affe9/$VENV_VERSION.tar.gz | tar xfz -
-
-$VENV_VERSION/virtualenv.py --system-site-packages --no-setuptools .env
-
-source .env/bin/activate
-
-curl -k https://bootstrap.pypa.io/ez_setup.py | python -
-curl -k https://gitlab.tiker.net/inducer/pip/raw/7.0.3/contrib/get-pip.py | python -
-
-pip install packaging
-
-PYTHON_VER=$(python -c 'import sys; print(".".join(str(s) for s in sys.version_info[:2]))')
-pip install git+https://github.com/pyinstaller/pyinstaller.git@413c37bec126c0bd26084813593f65128966b4b7
-
-git clone --recursive git://github.com/inducer/loopy
-cd loopy
-
-grep -v pyopencl requirements.txt > myreq.txt
-
-# needed for pyinstaller package to be usable
-echo packaging >> myreq.txt
-
-pip install -r myreq.txt
-python setup.py install
-
-chown -R user /tmp/build
-
-su user -p -c "cd /tmp/build && source .env/bin/activate && cd loopy && ./build-helpers/run-pyinstaller.sh"

build-helpers/make-linux-build-docker-inner.sh

-#! /bin/bash
-
-set -e
-set -x
-
-mkdir /tmp/build
-cd /tmp/build
-
-useradd -d /home/user -m -s /bin/bash user
-
-yum install -y centos-release-scl
-yum install -y git python27 python27-python-devel python27-numpy tar gcc gcc-c++ mercurial libffi-devel
-
-scl enable python27 /mnt/make-linux-build-docker-inner-part-2.sh
-

build-helpers/make-linux-build-docker.sh

-#! /bin/bash
-
-# should be run in this directory (build-helpers)
-
-if test "$1" = "--nodate"; then
-  TGT_NAME=loopy-centos6
-else
-  TGT_NAME=loopy-centos6-$(date +"%Y-%m-%d")
-fi
-
-echo "Generating $TGT_NAME..."
-
-set -e
-set -x
-
-docker pull centos:6
-
-CNT=$(docker create -t -v $(pwd):/mnt centos:6 /mnt/make-linux-build-docker-inner.sh)
-echo "working in container $CNT"
-
-docker start -i $CNT
-
-docker cp $CNT:/tmp/build/loopy/dist/loopy $(pwd) || true
-
-mv loopy $TGT_NAME
-
-docker rm $CNT
-

build-helpers/run-pyinstaller.sh

-#! /bin/bash
-
-# run this from the loopy root directory
-
-rm -Rf dist build
-
-pyinstaller \
-  --workpath=build/pyinstaller \
-  build-helpers/loopy.spec

build-helpers/upload.sh

-#! /bin/bash
-
-set -e
-
-scp "$1" tiker.net:public_html/pub/loopy-binaries/

doc/index.rst

@@ -20,29 +20,6 @@ When you run this script, the following kernel is generated, compiled, and execu
 
 .. _static-binary:
 
-Want to try out loopy?
-----------------------
-
-There's no need to go through :ref:`installation` if you'd just like to get a
-feel for what loopy is.  Instead, you may
-`download a self-contained Linux binary <https://gitlab.tiker.net/inducer/loopy/-/jobs/66778/artifacts/browse/build-helpers/>`_.
-This is purposefully built on an ancient Linux distribution, so it should work
-on most versions of Linux that are currently out there.
-
-Once you have the binary, do the following::
-
-    chmod +x ./loopy-centos6
-    ./loopy-centos6 --target=opencl hello-loopy.loopy
-    ./loopy-centos6 --target=cuda hello-loopy.loopy
-    ./loopy-centos6 --target=ispc hello-loopy.loopy
-
-Grab the example here: :download:`examples/python/hello-loopy.loopy <../examples/python/hello-loopy.loopy>`.
-
-You may also donwload the most recent version by going to the `list of builds
-<https://gitlab.tiker.net/inducer/loopy/builds>`_, clicking on the newest one
-of type "CentOS binary", clicking on "Browse" under "Build Artifacts", then
-navigating to "build-helpers", and downloading the binary from there.
-
 Places on the web related to Loopy
 ----------------------------------
 

doc/ref_kernel.rst

@@ -151,6 +151,42 @@ Tag                             Meaning
 
 .. }}}
 
+Identifiers
+-----------
+
+Reserved Identifiers
+^^^^^^^^^^^^^^^^^^^^
+
+The identifier prefix ``_lp_`` is reserved for internal usage; when creating
+*inames*, *argument names*, *temporary variable names*, *substitution rule
+names*, *instruction IDs*, and other identifiers, users should *not* use names
+beginning with ``_lp_``.  This prefix is used for identifiers created
+internally when operating on Loopy's kernel IR. For Loopy developers, further
+information on name prefixes used within submodules is below.
+
+Identifier Registry
+^^^^^^^^^^^^^^^^^^^
+
+Functionality in :mod:`loopy` *must* use identifiers beginning with ``_lp_`` for
+all internally-created identifiers. Additionally, each name beginning with
+``_lp_`` must start with one of the reserved prefixes below. New prefixes may
+be registered by adding them to the table below. New prefixes may not themselves
+be the prefix of an existing prefix.
+
+**Reserved Identifier Prefixes**
+
+======================= ==================================
+Reserved Prefix         Usage (module or purpose)
+======================= ==================================
+``_lp_linchk_``         :mod:`loopy.linearization.checker`
+======================= ==================================
+
+.. note::
+
+    Existing Loopy code may not yet fully satisfy these naming requirements.
+    Name changes are in progress, and prefixes will be added to this registry
+    as they are created.
+
 .. _instructions:
 
 Instructions

doc/ref_transform.rst

@@ -118,7 +118,7 @@ Finishing up
 
 .. autofunction:: generate_loop_schedules
 
-.. autofunction:: get_one_scheduled_kernel
+.. autofunction:: get_one_linearized_kernel
 
 .. autofunction:: save_and_reload_temporaries
 

doc/tutorial.rst

@@ -1204,16 +1204,16 @@ Here is what happens when we try to generate code for the kernel:
 
 This happens due to the kernel splitting done by :mod:`loopy`. The splitting
 happens when the instruction schedule is generated. To see the schedule, we
-should call :func:`loopy.get_one_scheduled_kernel`:
+should call :func:`loopy.get_one_linearized_kernel`:
 
-   >>> knl = lp.get_one_scheduled_kernel(lp.preprocess_kernel(knl))
+   >>> knl = lp.get_one_linearized_kernel(lp.preprocess_kernel(knl))
    >>> print(knl)
    ---------------------------------------------------------------------------
    KERNEL: rotate_v2
    ---------------------------------------------------------------------------
    ...
    ---------------------------------------------------------------------------
-   SCHEDULE:
+   LINEARIZATION:
       0: CALL KERNEL rotate_v2(extra_args=[], extra_inames=[])
       1:     tmp = arr[i_inner + i_outer*16]  {id=maketmp}
       2: RETURN FROM KERNEL rotate_v2
@@ -1233,12 +1233,12 @@ goes for local temporaries).
 :func:`loopy.save_and_reload_temporaries` for the purpose of handling the
 task of saving and restoring temporary values across global barriers. This
 function adds instructions to the kernel without scheduling them. That means
-that :func:`loopy.get_one_scheduled_kernel` needs to be called one more time to
+that :func:`loopy.get_one_linearized_kernel` needs to be called one more time to
 put those instructions into the schedule.
 
-   >>> knl = lp.get_one_scheduled_kernel(lp.preprocess_kernel(knl))
+   >>> knl = lp.get_one_linearized_kernel(lp.preprocess_kernel(knl))
    >>> knl = lp.save_and_reload_temporaries(knl)
-   >>> knl = lp.get_one_scheduled_kernel(knl)  # Schedule added instructions
+   >>> knl = lp.get_one_linearized_kernel(knl)  # Schedule added instructions
    >>> print(knl)
    ---------------------------------------------------------------------------
    KERNEL: rotate_v2
@@ -1251,7 +1251,7 @@ put those instructions into the schedule.
    ---------------------------------------------------------------------------
    ...
    ---------------------------------------------------------------------------
-   SCHEDULE:
+   LINEARIZATION:
       0: CALL KERNEL rotate_v2(extra_args=['tmp_save_slot'], extra_inames=[])
       1:     tmp = arr[i_inner + i_outer*16]  {id=maketmp}
       2:     tmp_save_slot[tmp_save_hw_dim_0_rotate_v2, tmp_save_hw_dim_1_rotate_v2] = tmp  {id=tmp.save}

examples/fortran/matmul-driver.py

+import numpy as np
+import numpy.linalg as la
+import pyopencl as cl
+import pyopencl.array
+import pyopencl.clrandom
+import loopy as lp
+
+
+def main():
+    fn = "matmul.floopy"
+    with open(fn, "r") as inf:
+        source = inf.read()
+
+    dgemm, = lp.parse_transformed_fortran(source, filename=fn)
+
+    ctx = cl.create_some_context()
+    queue = cl.CommandQueue(ctx)
+
+    n = 2048
+    a = cl.array.empty(queue, (n, n), dtype=np.float64, order="F")
+    b = cl.array.empty(queue, (n, n), dtype=np.float64, order="F")
+    c = cl.array.zeros(queue, (n, n), dtype=np.float64, order="F")
+    cl.clrandom.fill_rand(a)
+    cl.clrandom.fill_rand(b)
+
+    dgemm = lp.set_options(dgemm, write_code=True)
+
+    dgemm(queue, a=a, b=b, alpha=1, c=c)
+
+    c_ref = (a.get() @ b.get())
+    assert la.norm(c_ref - c.get())/la.norm(c_ref) < 1e-10
+
+
+if __name__ == "__main__":
+    main()

loopy/__init__.py

@@ -123,12 +123,12 @@ from loopy.transform.add_barrier import add_barrier
 
 from loopy.type_inference import infer_unknown_types
 from loopy.preprocess import preprocess_kernel, realize_reduction
-from loopy.schedule import generate_loop_schedules, get_one_scheduled_kernel
-from loopy.statistics import (ToCountMap, CountGranularity, stringify_stats_mapping,
-        Op, MemAccess, get_op_poly, get_op_map, get_lmem_access_poly,
-        get_DRAM_access_poly, get_gmem_access_poly, get_mem_access_map,
-        get_synchronization_poly, get_synchronization_map,
-        gather_access_footprints, gather_access_footprint_bytes)
+from loopy.schedule import (
+    generate_loop_schedules, get_one_scheduled_kernel, get_one_linearized_kernel)
+from loopy.statistics import (ToCountMap, CountGranularity,
+        stringify_stats_mapping, Op, MemAccess, get_op_map, get_mem_access_map,
+        get_synchronization_map, gather_access_footprints,
+        gather_access_footprint_bytes)
 from loopy.codegen import (
         PreambleInfo,
         generate_code, generate_code_v2, generate_body)
@@ -248,16 +248,16 @@ __all__ = [
         "infer_unknown_types",
 
         "preprocess_kernel", "realize_reduction",
-        "generate_loop_schedules", "get_one_scheduled_kernel",
+        "generate_loop_schedules",
+        "get_one_scheduled_kernel", "get_one_linearized_kernel",
         "GeneratedProgram", "CodeGenerationResult",
         "PreambleInfo",
         "generate_code", "generate_code_v2", "generate_body",
 
         "ToCountMap", "CountGranularity", "stringify_stats_mapping", "Op",
-        "MemAccess", "get_op_poly", "get_op_map", "get_lmem_access_poly",
-        "get_DRAM_access_poly", "get_gmem_access_poly", "get_mem_access_map",
-        "get_synchronization_poly", "get_synchronization_map",
-        "gather_access_footprints", "gather_access_footprint_bytes",
+        "MemAccess",  "get_op_map", "get_mem_access_map",
+        "get_synchronization_map", "gather_access_footprints",
+        "gather_access_footprint_bytes",
 
         "CompiledKernel",
 

loopy/auto_test.py

@@ -534,7 +534,7 @@ def auto_test_vs_ref(
     from loopy.target.pyopencl import PyOpenCLTarget
     if test_knl.state not in [
             KernelState.PREPROCESSED,
-            KernelState.SCHEDULED]:
+            KernelState.LINEARIZED]:
         if isinstance(test_knl.target, PyOpenCLTarget):
             test_knl = test_knl.copy(target=PyOpenCLTarget(ctx.devices[0]))
 

loopy/check.py

@@ -184,6 +184,19 @@ def check_for_inactive_iname_access(kernel):
                         ", ".join(expression_inames - kernel.insn_inames(insn))))
 
 
+def check_for_unused_inames(kernel):
+    # Warn if kernel has unused inames
+    from loopy.transform.iname import get_used_inames
+    unused_inames = kernel.all_inames() - get_used_inames(kernel)
+    if unused_inames:
+        warn_with_kernel(
+            kernel, "unused_inames",
+            "Found unused inames in kernel: %s "
+            "Unused inames during linearization will be prohibited in "
+            "Loopy version 2021.X."
+            % unused_inames)
+
+
 def _is_racing_iname_tag(tv, tag):
     from loopy.kernel.data import (AddressSpace,
             LocalIndexTagBase, GroupIndexTag, ConcurrentTag, auto)
@@ -220,12 +233,12 @@ def check_for_write_races(kernel):
             assignee_inames = assignee_indices & kernel.all_inames()
             if not assignee_inames <= kernel.insn_inames(insn):
                 raise LoopyError(
-                        "assignee of instructiosn '%s' references "
+                        "assignee of instructions '%s' references "
                         "iname that the instruction does not depend on"
                         % insn.id)
 
             if assignee_name in kernel.arg_dict:
-                # Any parallel tags that are not depended upon by the assignee
+                # Any concurrent tags that are not depended upon by the assignee
                 # will cause write races.
 
                 raceable_parallel_insn_inames = set(
@@ -658,6 +671,7 @@ def pre_schedule_checks(kernel):
         check_loop_priority_inames_known(kernel)
         check_multiple_tags_allowed(kernel)
         check_for_inactive_iname_access(kernel)
+        check_for_unused_inames(kernel)
         check_for_write_races(kernel)
         check_for_data_dependent_parallel_bounds(kernel)
         check_bounds(kernel)