Skip to content
Snippets Groups Projects

Compare revisions

Changes are shown as if the source revision was being merged into the target revision. Learn more about comparing revisions.

Source

Select target project
No results found
Select Git revision
  • 66-absolute-access-is-probably-not-copied-correctly-after-_eval_subs
  • const_fix
  • fhennig/cpu-jit
  • fhennig/v2.0-deprecations
  • fma
  • gpu_bufferfield_fix
  • gpu_liveness_opts
  • holzer-master-patch-46757
  • hyteg
  • improved_comm
  • master
  • rangersbach/fix-gpu-atomics
  • target_dh_refactoring
  • v2.0-dev
  • vectorization_sqrt_fix
  • zikeliml/124-rework-tutorials
  • zikeliml/Task-96-dotExporterForAST
  • last/Kerncraft
  • last/LLVM
  • last/OpenCL
  • release/0.2.1
  • release/0.2.10
  • release/0.2.11
  • release/0.2.12
  • release/0.2.13
  • release/0.2.14
  • release/0.2.15
  • release/0.2.2
  • release/0.2.3
  • release/0.2.4
  • release/0.2.6
  • release/0.2.7
  • release/0.2.8
  • release/0.2.9
  • release/0.3.0
  • release/0.3.1
  • release/0.3.2
  • release/0.3.3
  • release/0.3.4
  • release/0.4.0
  • release/0.4.1
  • release/0.4.2
  • release/0.4.3
  • release/0.4.4
  • release/1.0
  • release/1.0.1
  • release/1.1
  • release/1.1.1
  • release/1.2
  • release/1.3
  • release/1.3.1
  • release/1.3.2
  • release/1.3.3
  • release/1.3.4
  • release/1.3.5
  • release/1.3.6
  • release/1.3.7
  • release/2.0.dev0
58 results

Target

Select target project
  • anirudh.jonnalagadda/pystencils
  • hyteg/pystencils
  • jbadwaik/pystencils
  • jngrad/pystencils
  • itischler/pystencils
  • ob28imeq/pystencils
  • hoenig/pystencils
  • Bindgen/pystencils
  • hammer/pystencils
  • da15siwa/pystencils
  • holzer/pystencils
  • alexander.reinauer/pystencils
  • ec93ujoh/pystencils
  • Harke/pystencils
  • seitz/pystencils
  • pycodegen/pystencils
16 results
Select Git revision
Show changes
Commits on Source (1072)
[flake8] [flake8]
max-line-length=120 max-line-length=120
exclude=pystencils/jupytersetup.py, exclude=src/pystencils/jupyter.py,
pystencils/plot2d.py src/pystencils/plot.py
pystencils/session.py src/pystencils/session.py
ignore = W293 W503 W291 ignore = W293 W503 W291 C901 E741
src/pystencils/_version.py export-subst
__pycache__ __pycache__
.ipynb_checkpoints .ipynb_checkpoints
.coverage .coverage*
*.pyc *.pyc
*.vti *.vti
/build /build
/dist /dist
/*.egg-info *.egg-info
.cache .cache
_build _build
/html_doc
/.idea /.idea
.vscode
.cache .cache
_local_tmp _local_tmp
\ No newline at end of file RELEASE-VERSION
test-report
src/pystencils/boundaries/createindexlistcython.c
src/pystencils/boundaries/createindexlistcython.*.so
tests/tmp
tests/var
tests/kerncraft_inputs/.2d-5pt.c_kerncraft/
tests/kerncraft_inputs/.3d-7pt.c_kerncraft/
report.xml
coverage_report/
# macOS
**/.DS_Store
*.uuid
stages: stages:
- pretest
- test - test
- nightly
- docs
- deploy - deploy
# -------------------------- Templates ------------------------------------------------------------------------------------
# Base configuration for jobs meant to run at every commit
.every-commit:
rules:
- if: $CI_PIPELINE_SOURCE != "schedule"
# Configuration for jobs meant to run on each commit to pycodegen/pystencils/master
.every-commit-master:
rules:
- if: '$CI_PIPELINE_SOURCE != "schedule" && $CI_PROJECT_PATH == "pycodegen/pystencils" && $CI_COMMIT_BRANCH == "master"'
# Base configuration for jobs meant to run at a schedule
.scheduled:
rules:
- if: $CI_PIPELINE_SOURCE == "schedule"
# -------------------------- Tests ------------------------------------------------------------------------------------ # -------------------------- Tests ------------------------------------------------------------------------------------
# Normal test - runs on every commit all but "long run" tests # Normal test - runs on every commit all but "long run" tests
tests-and-coverage: tests-and-coverage:
stage: test stage: pretest
except: extends: .every-commit
variables: image: i10git.cs.fau.de:5005/pycodegen/pycodegen/full:cupy12.3
- $ENABLE_NIGHTLY_BUILDS before_script:
image: i10git.cs.fau.de:5005/software/pystencils/full - pip install -e .
script: script:
- env
- pip list
- export NUM_CORES=$(nproc --all) - export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib - mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc - echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- mkdir public - mkdir public
- py.test -v -n $NUM_CORES --cov-report html --cov-report term --cov=. -m "not longrun" - pytest -v -n $NUM_CORES --cov-report html --cov-report xml --cov-report term --cov=. -m "not longrun" --html test-report/index.html --junitxml=report.xml
- python -m coverage xml
tags: tags:
- docker - docker
- cuda - cuda11
- AVX - AVX
coverage: /Total coverage:\s\d+.\d+\%/
artifacts: artifacts:
when: always when: always
paths: paths:
- coverage_report - coverage_report
- test-report
reports:
coverage_report:
coverage_format: cobertura
path: coverage.xml
junit: report.xml
# Nightly test - runs "long run" jobs only # Normal test with longruns
test-longrun: tests-and-coverage-with-longrun:
stage: test stage: test
only: when: manual
variables: allow_failure: true
- $ENABLE_NIGHTLY_BUILDS image: i10git.cs.fau.de:5005/pycodegen/pycodegen/full:cupy12.3
image: i10git.cs.fau.de:5005/software/pystencils/full before_script:
- pip install sympy --upgrade
- pip install -e .
script: script:
- env
- pip list
- export NUM_CORES=$(nproc --all) - export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib - mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc - echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- py.test -v -n $NUM_CORES --cov-report html --cov-report term --cov=. - mkdir public
- py.test -v -n $NUM_CORES
tags:
- docker
- cuda11
- AVX
# pipeline with latest python version
latest-python:
stage: test
extends: .every-commit
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/latest_python
before_script:
- pip install -e .
script:
- env
- pip list
- pip install -e .
- export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- mkdir public
- py.test -v -n $NUM_CORES -m "not longrun" --junitxml=report.xml
tags: tags:
- docker - docker
- cuda
- AVX - AVX
artifacts: artifacts:
paths: when: always
- coverage_report reports:
junit: report.xml
# Minimal tests in windows environment # Minimal tests in windows environment
minimal-windows: #minimal-windows:
# stage: test
# tags:
# - win
# script:
# - export NUM_CORES=$(nproc --all)
# - source /cygdrive/c/Users/build/Miniconda3/Scripts/activate
# - source activate pystencils
# - pip install joblib
# - pip list
# - python -c "import numpy"
# - py.test -v -m "not (notebook or longrun)"
ubuntu:
stage: test stage: test
except: extends: .every-commit
variables: image: i10git.cs.fau.de:5005/pycodegen/pycodegen/ubuntu
- $ENABLE_NIGHTLY_BUILDS before_script:
tags: - ln -s /usr/include/locale.h /usr/include/xlocale.h
- win - pip3 install -e .
script: script:
- source /cygdrive/c/Users/build/Miniconda3/Scripts/activate - export NUM_CORES=$(nproc --all)
- source activate pystencils_dev - mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- sed -i 's/--doctest-modules //g' pytest.ini
- env - env
- conda env list - pip list
- python -c "import numpy" - pytest -v -n $NUM_CORES -m "not longrun" --junitxml=report.xml
- python setup.py quicktest tags:
- docker
- cuda11
- AVX
artifacts:
when: always
reports:
junit: report.xml
minimal-ubuntu: .multiarch_template:
stage: test stage: test
except: extends: .every-commit
variables: allow_failure: true
- $ENABLE_NIGHTLY_BUILDS before_script: &multiarch_before_script
image: i10git.cs.fau.de:5005/software/pystencils/minimal_ubuntu # - pip3 install -v .
- export PYTHONPATH=src
- python3 -c "import pystencils as ps; ps.cpu.cpujit.read_config()"
- sed -i '/^fail_under.*/d' pytest.ini
script: script:
- python3 setup.py quicktest - export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- sed -i 's/--doctest-modules //g' pytest.ini
- env
- pip3 list
- python3 -m pytest -v -n $NUM_CORES --cov-report html --cov-report xml --cov=. --junitxml=report.xml tests/test_*vec*.py tests/test_random.py tests/test_half_precision.py
- python3 -m coverage xml
tags: tags:
- docker - docker
- AVX
artifacts:
when: always
paths:
- coverage_report
reports:
coverage_report:
coverage_format: cobertura
path: coverage.xml
junit: report.xml
arm64v8:
extends: .multiarch_template
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/arm64
variables:
QEMU_CPU: "cortex-a76"
before_script:
- *multiarch_before_script
- sed -i s/march=native/march=armv8-a/g ~/.config/pystencils/config.json
ppc64le:
extends: .multiarch_template
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/ppc64le
before_script:
- *multiarch_before_script
- sed -i s/mcpu=native/mcpu=power8/g ~/.config/pystencils/config.json
arm64v9:
# SVE support is still unreliable in GCC 11 (incorrect code for fixed-width vectors, internal compiler errors).
# For half precision Clang is necessary
extends: .multiarch_template
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/arm64
before_script:
- *multiarch_before_script
- sed -i s/march=native/march=armv9-a+sve2+sme/g ~/.config/pystencils/config.json
- sed -i s/g\+\+/clang++/g ~/.config/pystencils/config.json
riscv64:
# RISC-V vector extension are currently not supported by GCC.
extends: .multiarch_template
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/riscv64
variables:
# explicitly set SIMD as detection requires QEMU >= 8.1
PYSTENCILS_SIMD: "rvv"
QEMU_CPU: "rv64,v=true,zicboz=true"
before_script:
- *multiarch_before_script
- sed -i 's/march=native/march=rv64imfdvzicboz/g' ~/.config/pystencils/config.json
- sed -i s/g\+\+/clang++-15/g ~/.config/pystencils/config.json
minimal-conda: minimal-conda:
stage: pretest
extends: .every-commit
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/minimal_conda
before_script:
- pip install -e .
script:
- python quicktest.py
tags:
- docker
- cuda
minimal-sympy-master:
stage: test stage: test
except: extends: .every-commit
variables: image: i10git.cs.fau.de:5005/pycodegen/pycodegen/minimal_conda
- $ENABLE_NIGHTLY_BUILDS before_script:
image: i10git.cs.fau.de:5005/software/pystencils/minimal_conda - pip install -e .
script: script:
- python setup.py quicktest - python -m pip install --upgrade git+https://github.com/sympy/sympy.git
- python quicktest.py
allow_failure: true
tags: tags:
- docker - docker
- cuda
pycodegen-integration:
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/full
stage: test
when: manual
allow_failure: true
script:
- git clone https://gitlab-ci-token:${CI_JOB_TOKEN}@i10git.cs.fau.de/pycodegen/pycodegen.git
- cd pycodegen
- git submodule sync --recursive
- git submodule update --init --recursive
- git submodule foreach git fetch origin # compare the latest master version!
- git submodule foreach git reset --hard origin/master
- cd pystencils
- git remote add test $CI_REPOSITORY_URL
- git fetch test
- git reset --hard $CI_COMMIT_SHA
- cd ..
- pip install -e pystencils/
- pip install -e lbmpy/
- cmake --version
- ./install_walberla.sh
- export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- cd pystencils
- py.test -v -n $NUM_CORES --junitxml=report.xml .
- cd ../lbmpy
- py.test -v -n $NUM_CORES --junitxml=report.xml .
- cd ../walberla/build/
- make -j $NUM_CORES CodegenJacobiCPU CodegenJacobiGPU CodegenPoissonCPU CodegenPoissonGPU MicroBenchmarkGpuLbm LbCodeGenerationExample
- make -j $NUM_CORES multiphaseCPU multiphaseGPU FluctuatingMRT FlowAroundSphereCodeGen FieldLayoutAndVectorizationTest GeneratedOutflowBC
- cd apps/benchmarks/UniformGridGPU
- make -j $NUM_CORES
- cd ../UniformGridCPU
- make -j $NUM_CORES
tags:
- docker
- cuda11
- AVX
artifacts:
when: always
reports:
junit: pycodegen/*/report.xml
# -------------------- Scheduled Tasks --------------------------------------------------------------------------
# Nightly test against the latest (pre-release) version of SymPy published on PyPI
nightly-sympy:
stage: nightly
needs: []
extends: .scheduled
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/latest_python
before_script:
- pip install -e .
- pip install --upgrade --pre sympy
script:
- env
- pip list
- export NUM_CORES=$(nproc --all)
- mkdir -p ~/.config/matplotlib
- echo "backend:template" > ~/.config/matplotlib/matplotlibrc
- mkdir public
- pytest -v -n $NUM_CORES -m "not longrun" --junitxml=report.xml
tags:
- docker
- AVX
- cuda
artifacts:
when: always
reports:
junit: report.xml
# -------------------- Linter & Documentation -------------------------------------------------------------------------- # -------------------- Linter & Documentation --------------------------------------------------------------------------
flake8-lint: flake8-lint:
stage: test stage: pretest
except: extends: .every-commit
variables: image: i10git.cs.fau.de:5005/pycodegen/pycodegen/full
- $ENABLE_NIGHTLY_BUILDS
image: i10git.cs.fau.de:5005/software/pystencils/full
script: script:
- flake8 pystencils - flake8 src/pystencils
tags: tags:
- docker - docker
- cuda
build-documentation: build-documentation:
stage: test stage: docs
image: i10git.cs.fau.de:5005/software/pystencils/full extends: .every-commit
image: i10git.cs.fau.de:5005/pycodegen/pycodegen/documentation
needs: []
before_script:
- pip install -e .
script: script:
- export PYTHONPATH=`pwd`
- mkdir html_doc - mkdir html_doc
- sphinx-build -b html doc html_doc
- sphinx-build -W -b html doc html_doc - sphinx-build -W -b html doc html_doc
tags: tags:
- docker - docker
- cuda
artifacts: artifacts:
paths: paths:
- html_doc - html_doc
pages: pages:
image: i10git.cs.fau.de:5005/software/pystencils/full image: i10git.cs.fau.de:5005/pycodegen/pycodegen/full
extends: .every-commit-master
stage: deploy stage: deploy
needs: ["tests-and-coverage", "build-documentation"]
script: script:
- ls -l - ls -l
- mv coverage_report html_doc - mv coverage_report html_doc
...@@ -129,5 +359,3 @@ pages: ...@@ -129,5 +359,3 @@ pages:
- public - public
tags: tags:
- docker - docker
only:
- master@pycodegen/pystencils
[settings]
line_length=100
balanced_wrapping=True
multi_line_output=4
known_third_party=sympy
Contributors:
-------------
- Martin Bauer <martin.bauer@fau.de>
- Markus Holzer <markus.holzer@fau.de>
- Stephan Seitz <stephan.seitz@fau.de>
- Michael Kuron <mkuron@icp.uni-stuttgart.de>
- Jan Hönig <jan.hoenig@fau.de>
- Julian Hammer <julian.hammer@fau.de>
- Nils Kohl <nils.kohl@fau.de>
- Frederik Hennig <frederik.hennig@fau.de>
- Dominik Ernst <dominik.ernst@fau.de>
- Christian Godenschwager <christian.godenschwager@fau.de>
- Dominik Thoennes <dominik.thoennes@fau.de>
# Change Log
## Unreleased
### Removed
* LLVM backend because it was not used much and not good integrated in pystencils.
* OpenCL backend because it was not used much and not good integrated in pystencils.
# Contributing
Contributions to pystencils are always welcome, and they are greatly appreciated!
A list of open problems can be found [here]( https://i10git.cs.fau.de/pycodegen/pystencils/-/issues).
Of course, it is also always appreciated to bring own ideas and problems to the community!
Please submit all contributions to the official [GitLab repository](https://i10git.cs.fau.de/pycodegen/pystencils) in the form of a Merge Request. Please do not submit git diffs or files containing the changes.
There also exists a GitHub repository, which is only a mirror to the GitLab repository. Contributions to the GitHub repository are not considered.
`pystencils` is an open-source python package under the license of AGPL3. Thus we consider the act of contributing to the code by submitting a Merge Request as the "Sign off" or agreement to the AGPL3 license.
You can contribute in many different ways:
## Types of Contributions
### Report Bugs
Report bugs at [https://i10git.cs.fau.de/pycodegen/pystencils/-/issues](https://i10git.cs.fau.de/pycodegen/pystencils/-/issues).
For pystencils, it is often necessary to provide the python and [SymPy](https://www.sympy.org/en/index.html) versions used and hardware information like the
processor architecture and the compiler version used to compile the generated kernels.
### Fix Issues
Look through the GitLab issues. Different tags are indicating the status of the issues.
The "bug" tag indicates problems with pystencils, while the "feature" tag shows ideas that should be added in the future.
### Write Documentation
The documentation of pystencils can be found [here](https://pycodegen.pages.i10git.cs.fau.de/pystencils/). Jupyter notebooks are used to provide an
interactive start to pystencils. It is always appreciated if new document notebooks are provided
since this helps others a lot.
## Get Started!
Ready to contribute? Here is how to set up `pystencils` for local development.
1. Fork the `pystencils` repo on GitLab.
2. Clone your fork locally:
```bash
$ git clone https://i10git.cs.fau.de/your-name/pystencils
```
3. Install your local copy into a virtualenv. It is also recommended to use anaconda or miniconda to manage the python environments.
```bash
$ mkvirtualenv pystencils
$ cd pystencils/
$ pip install -e .
```
4. Create a branch for local development:
```bash
$ git checkout -b name-of-your-bugfix-or-feature
```
Now you can make your changes locally.
5. When you're done making changes, check that your changes pass flake8 and the
tests
```bash
$ flake8 pystencils
$ py.test -v -n $NUM_CORES -m "not longrun" .
```
To get all packages needed for development, a requirements list can be found [here](https://i10git.cs.fau.de/pycodegen/pycodegen/-/blob/master/conda_environment_dev.yml). This includes flake8 and pytest.
6. Commit your changes and push your branch to GitHub::
```bash
$ git add .
$ git commit -m "Your detailed description of your changes."
$ git push origin name-of-your-bugfix-or-feature
```
7. Submit a Merge Request on GitLab.
## Merge Request Guidelines
Before you submit a Merge Request, check that it meets these guidelines:
1. All functionality that is implemented through this Merge Request should be covered by unit tests. These are implemented in `pystencil_tests`
2. If the Merge Request adds functionality, the docs should be updated. Put your new functionality into a function with a docstring.
3. If you have a maintainer status for `pystencils`, you can merge Merge Requests to the master branch. However, every Merge Request needs to be reviewed by another developer. Thus it is not allowed to merge a Merge Request, which is submitted by oneself.
## Tips
To run a subset of tests:
```bash
$ py.test my_test.py
```
\ No newline at end of file
include README.md include AUTHORS.txt
include COPYING.txt include CONTRIBUTING.md
include CHANGELOG.md
pystencils pystencils
========== ==========
[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/mabau/pystencils/master?filepath=doc%2Fnotebooks)
[![Docs](https://img.shields.io/badge/read-the_docs-brightgreen.svg)](https://pycodegen.pages.i10git.cs.fau.de/pystencils)
[![pypi-package](https://badge.fury.io/py/pystencils.svg)](https://badge.fury.io/py/pystencils)
[![pipeline status](https://i10git.cs.fau.de/pycodegen/pystencils/badges/master/pipeline.svg)](https://i10git.cs.fau.de/pycodegen/pystencils/commits/master) [![pipeline status](https://i10git.cs.fau.de/pycodegen/pystencils/badges/master/pipeline.svg)](https://i10git.cs.fau.de/pycodegen/pystencils/commits/master)
[![coverage report](https://i10git.cs.fau.de/pycodegen/pystencils/badges/master/coverage.svg)](https://i10git.cs.fau.de/pycodegen/pystencils/commits/master) [![coverage report](https://i10git.cs.fau.de/pycodegen/pystencils/badges/master/coverage.svg)](http://pycodegen.pages.i10git.cs.fau.de/pystencils/coverage_report)
[coverage report](http://pycodegen.pages.walberla.net/pystencils/coverage_report)
Run blazingly fast stencil codes on numpy arrays. Run blazingly fast stencil codes on numpy arrays.
*pystencils* uses sympy to define stencil operations, that can be executed on numpy array. *pystencils* uses sympy to define stencil operations, that can be executed on numpy arrays.
It runs faster than normal numpy code and even as Cython and numba. Exploiting the stencil structure makes *pystencils* run faster than normal numpy code and even as Cython and numba,
[as demonstrated in this notebook](https://pycodegen.pages.i10git.cs.fau.de/pystencils/notebooks/demo_benchmark.html).
Here is a code snippet that computes the average of neighboring cells: Here is a code snippet that computes the average of neighboring cells:
```python ```python
...@@ -17,7 +21,7 @@ import numpy as np ...@@ -17,7 +21,7 @@ import numpy as np
f, g = ps.fields("f, g : [2D]") f, g = ps.fields("f, g : [2D]")
stencil = ps.Assignment(g[0, 0], stencil = ps.Assignment(g[0, 0],
(f[1, 0] + f[-1, 0] + f[0, 1] + f[0, -1]) / 4) (f[1, 0] + f[-1, 0] + f[0, 1] + f[0, -1]) / 4)
kernel = ps.create_kernel(stencil).compile() kernel = ps.create_kernel(stencil).compile()
f_arr = np.random.rand(1000, 1000) f_arr = np.random.rand(1000, 1000)
...@@ -29,15 +33,15 @@ kernel(f=f_arr, g=g_arr) ...@@ -29,15 +33,15 @@ kernel(f=f_arr, g=g_arr)
It comes with automatic finite difference discretization for PDEs: It comes with automatic finite difference discretization for PDEs:
```python ```python
import pystencils as ps
import sympy as sp
c, v = ps.fields("c, v(2): [2D]") c, v = ps.fields("c, v(2): [2D]")
adv_diff_pde = ps.fd.transient(c) - ps.fd.diffusion(c, sp.symbols("D")) + ps.fd.advection(c, v) adv_diff_pde = ps.fd.transient(c) - ps.fd.diffusion(c, sp.symbols("D")) + ps.fd.advection(c, v)
discretize = ps.fd.Discretization2ndOrder(dx=1, dt=0.01) discretize = ps.fd.Discretization2ndOrder(dx=1, dt=0.01)
discretization = discretize(adv_diff_pde) discretization = discretize(adv_diff_pde)
``` ```
Look at the [documentation](http://pycodegen.pages.walberla.net/pystencils) to learn more.
Installation Installation
------------ ------------
...@@ -48,7 +52,7 @@ pip install pystencils[interactive] ...@@ -48,7 +52,7 @@ pip install pystencils[interactive]
Without `[interactive]` you get a minimal version with very little dependencies. Without `[interactive]` you get a minimal version with very little dependencies.
All options: All options:
- `gpu`: use this if nVidia GPU is available and CUDA is installed - `gpu`: use this if an NVIDIA or AMD GPU is available and CUDA or ROCm is installed
- `alltrafos`: pulls in additional dependencies for loop simplification e.g. libisl - `alltrafos`: pulls in additional dependencies for loop simplification e.g. libisl
- `bench_db`: functionality to store benchmark result in object databases - `bench_db`: functionality to store benchmark result in object databases
- `interactive`: installs dependencies to work in Jupyter including image I/O, plotting etc. - `interactive`: installs dependencies to work in Jupyter including image I/O, plotting etc.
...@@ -56,12 +60,28 @@ All options: ...@@ -56,12 +60,28 @@ All options:
Options can be combined e.g. Options can be combined e.g.
```bash ```bash
pip install pystencils[interactive,gpu,doc] pip install pystencils[interactive, gpu, doc]
``` ```
pystencils is also fully compatible with Windows machines. If working with visual studio and cupy makes sure to run example files first to ensure that cupy can find the compiler's executable.
Documentation Documentation
------------- -------------
Read the docs [here](http://pycodegen.pages.walberla.net/pystencils) and Read the docs [here](https://pycodegen.pages.i10git.cs.fau.de/pystencils) and
check out the Jupyter notebooks in `doc/notebooks`. check out the Jupyter notebooks in `doc/notebooks`. The **Changelog** of pystencils can be found [here](https://i10git.cs.fau.de/pycodegen/pystencils/-/blob/master/CHANGELOG.md).
Authors
-------
Many thanks go to the [contributors](https://i10git.cs.fau.de/pycodegen/pystencils/-/blob/master/AUTHORS.txt) of pystencils.
### Please cite us
If you use pystencils in a publication, please cite the following articles:
Overview:
- M. Bauer et al, Code Generation for Massively Parallel Phase-Field Simulations. Association for Computing Machinery, 2019. https://doi.org/10.1145/3295500.3356186
Performance Modelling:
- D. Ernst et al, Analytical performance estimation during code generation on modern GPUs. Journal of Parallel and Distributed Computing, 2023. https://doi.org/10.1016/j.jpdc.2022.11.003
build-essential
graphviz
ffmpeg
# ----------------------------------------------------------------------------------------------------------------------
# Environment with all dependencies to use pystencils
#
#
# Download conda at https://conda.io/miniconda.html and create this environment by running:
# conda env create -f conda_environment_user.yml
# . activate pystencils
#
# If you have CUDA or ROCm installed and want to use your GPU, uncomment the last line to install cupy
#
# ----------------------------------------------------------------------------------------------------------------------
name: pystencils
dependencies:
# Basic dependencies:
- python >= 3.8
- numpy
- sympy >= 1.1
- appdirs # to find default cache directory on each platform
- joblib # caching on hard-disk, this is optional, but if not installed lbmpy is really slow
- cython # speed up boundary list computation (optional)
- matplotlib
- imageio
- pandas
- scipy
- pip
- pip:
- islpy # used to optimize staggered kernels
- py-cpuinfo # get cpu info like cache sizes, supported vector instruction sets, ...
- graphviz # can show abstract syntax trees as formatted graphs
- ipy_table # HTML tables for jupyter notebooks
- pyevtk # VTK output for serial simulations
- blitzdb # file-based No-SQL database to store simulation results
#- cupy # add this if you have CUDA or ROCm installed
#!/bin/bash
python3 setup.py develop
exec "$@"
\ No newline at end of file
import os import os
import pytest
import tempfile
import runpy import runpy
import sys import sys
import tempfile
import warnings
import nbformat
import pytest
from nbconvert import PythonExporter
from pystencils.boundaries.createindexlist import * # NOQA
# Trigger config file reading / creation once - to avoid race conditions when multiple instances are creating it # Trigger config file reading / creation once - to avoid race conditions when multiple instances are creating it
# at the same time # at the same time
from pystencils.cpu import cpujit from pystencils.cpu import cpujit
...@@ -14,12 +20,17 @@ try: ...@@ -14,12 +20,17 @@ try:
pyximport.install(language_level=3) pyximport.install(language_level=3)
except ImportError: except ImportError:
pass pass
from pystencils.boundaries.createindexlistcython import * # NOQA
SCRIPT_FOLDER = os.path.dirname(os.path.realpath(__file__)) SCRIPT_FOLDER = os.path.dirname(os.path.realpath(__file__))
sys.path.insert(0, os.path.abspath('pystencils')) sys.path.insert(0, os.path.abspath('pystencils'))
# the Ubuntu pipeline uses an older version of pytest which uses deprecated functionality.
# This leads to many warinings in the test and coverage pipeline.
pytest_numeric_version = [int(x, 10) for x in pytest.__version__.split('.')]
pytest_numeric_version.reverse()
pytest_version = sum(x * (100 ** i) for i, x in enumerate(pytest_numeric_version))
def add_path_to_ignore(path): def add_path_to_ignore(path):
if not os.path.exists(path): if not os.path.exists(path):
...@@ -28,34 +39,49 @@ def add_path_to_ignore(path): ...@@ -28,34 +39,49 @@ def add_path_to_ignore(path):
collect_ignore += [os.path.join(SCRIPT_FOLDER, path, f) for f in os.listdir(os.path.join(SCRIPT_FOLDER, path))] collect_ignore += [os.path.join(SCRIPT_FOLDER, path, f) for f in os.listdir(os.path.join(SCRIPT_FOLDER, path))]
collect_ignore = [os.path.join(SCRIPT_FOLDER, "doc", "conf.py")] collect_ignore = [os.path.join(SCRIPT_FOLDER, "doc", "conf.py"),
add_path_to_ignore('pystencils_tests/benchmark') os.path.join(SCRIPT_FOLDER, "src", "pystencils", "opencl", "opencl.autoinit")]
add_path_to_ignore('tests/benchmark')
add_path_to_ignore('_local_tmp') add_path_to_ignore('_local_tmp')
try: try:
import pycuda import cupy
except ImportError: except ImportError:
collect_ignore += [os.path.join(SCRIPT_FOLDER, "pystencils/pystencils_tests/test_cudagpu.py")] collect_ignore += [os.path.join(SCRIPT_FOLDER, "tests/test_gpu.py")]
add_path_to_ignore('pystencils/gpucuda') add_path_to_ignore('src/pystencils/gpu')
try: try:
import llvmlite import waLBerla
except ImportError: except ImportError:
collect_ignore += [os.path.join(SCRIPT_FOLDER, 'pystencils_tests/backends/llvm.py')] collect_ignore += [os.path.join(SCRIPT_FOLDER, "tests/test_aligned_array.py"),
add_path_to_ignore('pystencils/llvm') os.path.join(SCRIPT_FOLDER, "tests/test_datahandling_parallel.py"),
os.path.join(SCRIPT_FOLDER, "doc/notebooks/03_tutorial_datahandling.ipynb"),
os.path.join(SCRIPT_FOLDER, "src/pystencils/datahandling/parallel_datahandling.py"),
os.path.join(SCRIPT_FOLDER, "tests/test_small_block_benchmark.ipynb")]
try: try:
import kerncraft import blitzdb
except ImportError: except ImportError:
collect_ignore += [os.path.join(SCRIPT_FOLDER, "pystencils_tests/test_kerncraft_coupling.py")] add_path_to_ignore('src/pystencils/runhelper')
add_path_to_ignore('pystencils/kerncraft_coupling') collect_ignore += [os.path.join(SCRIPT_FOLDER, "tests/test_parameterstudy.py")]
collect_ignore += [os.path.join(SCRIPT_FOLDER, "tests/test_json_serializer.py")]
try: try:
import blitzdb import islpy
except ImportError: except ImportError:
add_path_to_ignore('pystencils/runhelper') collect_ignore += [os.path.join(SCRIPT_FOLDER, "src/pystencils/integer_set_analysis.py")]
try:
import graphviz
except ImportError:
collect_ignore += [os.path.join(SCRIPT_FOLDER, "src/pystencils/backends/dot.py")]
collect_ignore += [os.path.join(SCRIPT_FOLDER, "doc/notebooks/01_tutorial_getting_started.ipynb")]
try:
import pyevtk
except ImportError:
collect_ignore += [os.path.join(SCRIPT_FOLDER, "src/pystencils/datahandling/vtk.py")]
collect_ignore += [os.path.join(SCRIPT_FOLDER, 'setup.py')] collect_ignore += [os.path.join(SCRIPT_FOLDER, 'setup.py')]
...@@ -65,10 +91,6 @@ for root, sub_dirs, files in os.walk('.'): ...@@ -65,10 +91,6 @@ for root, sub_dirs, files in os.walk('.'):
collect_ignore.append(f) collect_ignore.append(f)
import nbformat
from nbconvert import PythonExporter
class IPythonMockup: class IPythonMockup:
def run_line_magic(self, *args, **kwargs): def run_line_magic(self, *args, **kwargs):
pass pass
...@@ -89,6 +111,7 @@ class IPyNbTest(pytest.Item): ...@@ -89,6 +111,7 @@ class IPyNbTest(pytest.Item):
self.code = code self.code = code
self.add_marker('notebook') self.add_marker('notebook')
@pytest.mark.filterwarnings("ignore:IPython.core.inputsplitter is deprecated")
def runtest(self): def runtest(self):
global_dict = {'get_ipython': lambda: IPythonMockup(), global_dict = {'get_ipython': lambda: IPythonMockup(),
'is_test_run': True} 'is_test_run': True}
...@@ -114,10 +137,16 @@ class IPyNbFile(pytest.File): ...@@ -114,10 +137,16 @@ class IPyNbFile(pytest.File):
exporter.exclude_markdown = True exporter.exclude_markdown = True
exporter.exclude_input_prompt = True exporter.exclude_input_prompt = True
notebook_contents = self.fspath.open() notebook_contents = self.fspath.open(encoding='utf-8')
notebook = nbformat.read(notebook_contents, 4)
code, _ = exporter.from_notebook_node(notebook) with warnings.catch_warnings():
yield IPyNbTest(self.name, self, code) warnings.filterwarnings("ignore", "IPython.core.inputsplitter is deprecated")
notebook = nbformat.read(notebook_contents, 4)
code, _ = exporter.from_notebook_node(notebook)
if pytest_version >= 50403:
yield IPyNbTest.from_parent(name=self.name, parent=self, code=code)
else:
yield IPyNbTest(self.name, self, code)
def teardown(self): def teardown(self):
pass pass
...@@ -126,4 +155,7 @@ class IPyNbFile(pytest.File): ...@@ -126,4 +155,7 @@ class IPyNbFile(pytest.File):
def pytest_collect_file(path, parent): def pytest_collect_file(path, parent):
glob_exprs = ["*demo*.ipynb", "*tutorial*.ipynb", "test_*.ipynb"] glob_exprs = ["*demo*.ipynb", "*tutorial*.ipynb", "test_*.ipynb"]
if any(path.fnmatch(g) for g in glob_exprs): if any(path.fnmatch(g) for g in glob_exprs):
return IPyNbFile(path, parent) if pytest_version >= 50403:
return IPyNbFile.from_parent(fspath=path, parent=parent)
else:
return IPyNbFile(path, parent)
...@@ -4,10 +4,11 @@ ...@@ -4,10 +4,11 @@
import datetime import datetime
import sphinx_rtd_theme import sphinx_rtd_theme
import os import os
import re
import sys import sys
sys.path.insert(0, os.path.abspath('.')) sys.path.insert(0, os.path.abspath('.'))
from version_from_git import version_number_from_git import pystencils
extensions = [ extensions = [
'sphinx.ext.autodoc', 'sphinx.ext.autodoc',
...@@ -25,11 +26,14 @@ templates_path = ['_templates'] ...@@ -25,11 +26,14 @@ templates_path = ['_templates']
source_suffix = '.rst' source_suffix = '.rst'
master_doc = 'index' master_doc = 'index'
copyright = '{}, Martin Bauer'.format(datetime.datetime.now().year) copyright = f'{datetime.datetime.now().year}, Martin Bauer, Markus Holzer, Frederik Hennig'
author = 'Martin Bauer' author = 'Martin Bauer, Markus Holzer, Frederik Hennig'
version = version_number_from_git() # The short X.Y version (including .devXXXX, rcX, b1 suffixes if present)
release = version_number_from_git() version = re.sub(r'(\d+\.\d+)\.\d+(.*)', r'\1\2', pystencils.__version__)
language = None version = re.sub(r'(\.dev\d+).*?$', r'\1', version)
# The full version, including alpha/beta/rc tags.
release = pystencils.__version__
language = 'en'
exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store', '**.ipynb_checkpoints'] exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store', '**.ipynb_checkpoints']
default_role = 'any' default_role = 'any'
pygments_style = 'sphinx' pygments_style = 'sphinx'
...@@ -47,13 +51,14 @@ nbsphinx_execute = 'never' ...@@ -47,13 +51,14 @@ nbsphinx_execute = 'never'
nbsphinx_codecell_lexer = 'python3' nbsphinx_codecell_lexer = 'python3'
# Example configuration for intersphinx: refer to the Python standard library. # Example configuration for intersphinx: refer to the Python standard library.
intersphinx_mapping = {'python': ('https://docs.python.org/3.6', None), intersphinx_mapping = {'python': ('https://docs.python.org/3.8', None),
'numpy': ('https://docs.scipy.org/doc/numpy/', None), 'numpy': ('https://docs.scipy.org/doc/numpy/', None),
'matplotlib': ('https://matplotlib.org/', None), 'matplotlib': ('https://matplotlib.org/', None),
'sympy': ('https://docs.sympy.org/latest/', None), 'sympy': ('https://docs.sympy.org/latest/', None),
} }
autodoc_member_order = 'bysource' autodoc_member_order = 'bysource'
bibtex_bibfiles = ['sphinx/pystencils.bib']
project = 'pystencils' project = 'pystencils'
html_logo = "img/logo.png" html_logo = 'img/logo.png'
doc/img/github_repo_card.png

78.4 KiB

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!-- Created with Inkscape (http://www.inkscape.org/) -->
<svg
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:cc="http://creativecommons.org/ns#"
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:svg="http://www.w3.org/2000/svg"
xmlns="http://www.w3.org/2000/svg"
xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd"
xmlns:inkscape="http://www.inkscape.org/namespaces/inkscape"
width="379.82614mm"
height="189.91307mm"
viewBox="0 0 1345.8407 672.92033"
id="svg2"
version="1.1"
inkscape:version="0.92.3 (2405546, 2018-03-11)"
sodipodi:docname="logo_large.svg"
inkscape:export-filename="/home/martin/code/pycodegen/pystencils/doc/img/github_repo_card.png"
inkscape:export-xdpi="85.599998"
inkscape:export-ydpi="85.599998">
<defs
id="defs4">
<inkscape:path-effect
effect="spiro"
id="path-effect4188"
is_visible="true" />
<inkscape:path-effect
effect="spiro"
id="path-effect4188-5"
is_visible="true" />
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4596">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4598" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4600" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4602" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4604" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4606" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4608">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4610" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4612" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4614" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4616" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4618" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4620">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4622" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4624" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4626" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4628" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4630" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4632">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4634" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4636" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4638" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4640" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4642" />
</filter>
<inkscape:path-effect
effect="spiro"
id="path-effect4188-7"
is_visible="true" />
<inkscape:path-effect
effect="spiro"
id="path-effect4188-5-6"
is_visible="true" />
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4596-6">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4598-6" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4600-9" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4602-1" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4604-4" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4606-3" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4620-1">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4622-1" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4624-4" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4626-8" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4628-5" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4630-7" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4632-1">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4634-9" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4636-8" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4638-7" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4640-6" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4642-5" />
</filter>
<filter
y="-0.25"
height="1.5"
inkscape:menu-tooltip="Darkens the edge with an inner blur and adds a flexible glow"
inkscape:menu="Shadows and Glows"
inkscape:label="Dark And Glow"
style="color-interpolation-filters:sRGB"
id="filter4608-0">
<feGaussianBlur
stdDeviation="5"
result="result6"
id="feGaussianBlur4610-2" />
<feComposite
result="result8"
in="SourceGraphic"
operator="atop"
in2="result6"
id="feComposite4612-5" />
<feComposite
result="result9"
operator="over"
in2="SourceAlpha"
in="result8"
id="feComposite4614-7" />
<feColorMatrix
values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 "
result="result10"
id="feColorMatrix4616-6" />
<feBlend
in="result10"
mode="normal"
in2="result6"
id="feBlend4618-9" />
</filter>
</defs>
<sodipodi:namedview
id="base"
pagecolor="#ffffff"
bordercolor="#666666"
borderopacity="1.0"
inkscape:pageopacity="0.0"
inkscape:pageshadow="2"
inkscape:zoom="0.70000001"
inkscape:cx="545.01294"
inkscape:cy="35.725386"
inkscape:document-units="px"
inkscape:current-layer="layer1"
showgrid="false"
inkscape:window-width="3840"
inkscape:window-height="2061"
inkscape:window-x="0"
inkscape:window-y="0"
inkscape:window-maximized="1"
fit-margin-top="0"
fit-margin-left="0"
fit-margin-right="0"
fit-margin-bottom="0">
<inkscape:grid
type="xygrid"
id="grid4176"
originx="267.20477"
originy="315.17846" />
</sodipodi:namedview>
<metadata
id="metadata7">
<rdf:RDF>
<cc:Work
rdf:about="">
<dc:format>image/svg+xml</dc:format>
<dc:type
rdf:resource="http://purl.org/dc/dcmitype/StillImage" />
<dc:title></dc:title>
</cc:Work>
</rdf:RDF>
</metadata>
<g
inkscape:label="Layer 1"
inkscape:groupmode="layer"
id="layer1"
transform="translate(267.20477,-694.6203)">
<text
xml:space="preserve"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:34.78659058px;line-height:125%;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, ';letter-spacing:0px;word-spacing:0px;fill:#252525;fill-opacity:1;stroke:none;stroke-width:4.34832382px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1"
x="159.99139"
y="964.43109"
id="text1392-1"
inkscape:export-xdpi="70.669998"
inkscape:export-ydpi="70.669998"><tspan
sodipodi:role="line"
id="tspan1390-1"
x="159.99139"
y="964.43109"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:156.53968811px;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, Bold';fill:#252525;fill-opacity:1;stroke-width:4.34832382px">pystencils</tspan></text>
<g
id="g9986"
transform="matrix(4.1201463,0,0,4.1201463,-399.75066,866.02979)"
inkscape:export-xdpi="70.669998"
inkscape:export-ydpi="70.669998">
<path
inkscape:connector-curvature="0"
inkscape:original-d="M 60.891002,27.333516 H 118.64865"
inkscape:path-effect="#path-effect4188-7"
id="path4186-6"
d="M 60.891002,27.333516 H 118.64865"
style="fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2.78799796;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:none;stroke-opacity:0.70388345" />
<path
sodipodi:nodetypes="cc"
inkscape:connector-curvature="0"
inkscape:original-d="M 89.922623,-0.47572315 C 31.237244,132.88729 89.846228,36.88339 89.846228,56.13594"
inkscape:path-effect="#path-effect4188-5-6"
id="path4186-3-9"
d="M 89.922623,-0.47572315 89.846228,56.13594"
style="fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2.78799796;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:none;stroke-opacity:0.70388345" />
<circle
transform="matrix(0.21391721,0,0,0.21391721,27.733834,-23.442344)"
r="34.345188"
cy="108.02044"
cx="291.42902"
id="path4136-76"
style="opacity:1;fill:#e69f00;fill-opacity:1;stroke:none;stroke-width:3;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;filter:url(#filter4596-6)" />
<circle
transform="matrix(0.21391721,0,0,0.21391721,27.733834,-23.442344)"
r="34.345188"
cy="365.43817"
cx="290.41885"
id="path4136-6-0"
style="opacity:1;fill:#0072b2;fill-opacity:1;stroke:none;stroke-width:3;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;filter:url(#filter4620-1)" />
<circle
transform="matrix(0.21391721,0,0,0.21391721,27.733834,-23.442344)"
r="34.345188"
cy="236.72931"
cx="422.24377"
id="path4136-3-9"
style="opacity:1;fill:#999999;fill-opacity:1;stroke:none;stroke-width:3;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;filter:url(#filter4632-1)" />
<circle
transform="matrix(0.21391721,0,0,0.21391721,27.733834,-23.442344)"
r="34.345188"
cy="236.72931"
cx="155.56349"
id="path4136-7-0"
style="opacity:1;fill:#009e73;fill-opacity:1;stroke:none;stroke-width:3;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;filter:url(#filter4608-0)" />
</g>
<text
xml:space="preserve"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:8.7668047px;line-height:125%;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, ';letter-spacing:0px;word-spacing:0px;fill:#252525;fill-opacity:0.70629368;stroke:none;stroke-width:1.09585059px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1"
x="229.96391"
y="1071.713"
id="text1392-1-3"
inkscape:export-xdpi="70.669998"
inkscape:export-ydpi="70.669998"><tspan
sodipodi:role="line"
id="tspan1390-1-6"
x="229.96391"
y="1071.713"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:62.0406723px;line-height:105.99999428%;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, Bold';fill:#252525;fill-opacity:0.70629368;stroke-width:1.09585059px">speed up stencil </tspan><tspan
sodipodi:role="line"
x="229.96391"
y="1137.4761"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:62.0406723px;line-height:105.99999428%;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, Bold';fill:#252525;fill-opacity:0.70629368;stroke-width:1.09585059px"
id="tspan109">computations on</tspan><tspan
sodipodi:role="line"
x="229.96391"
y="1203.2393"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:62.0406723px;line-height:105.99999428%;font-family:'Latin Modern Mono Light';-inkscape-font-specification:'Latin Modern Mono Light, Bold';fill:#252525;fill-opacity:0.70629368;stroke-width:1.09585059px"
id="tspan107">numpy arrays</tspan></text>
</g>
</svg>
...@@ -14,5 +14,6 @@ pystencils can help you to generate blazingly fast code for image processing, nu ...@@ -14,5 +14,6 @@ pystencils can help you to generate blazingly fast code for image processing, nu
.. image:: /img/pystencils_arch_block_diagram.svg .. image:: /img/pystencils_arch_block_diagram.svg
:height: 450px :height: 450px
:align: center
Source diff could not be displayed: it is too large. Options to address this: view the blob.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
from pystencils.session import * from pystencils.session import *
``` ```
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
# Tutorial 02: Basic Kernel generation with *pystencils* # Tutorial 02: Basic Kernel generation with *pystencils*
Now that you have an [overview of pystencils](01_tutorial_getting_started.ipynb), Now that you have an [overview of pystencils](01_tutorial_getting_started.ipynb),
this tutorial shows in more detail how to formulate, optimize and run stencil kernels. this tutorial shows in more detail how to formulate, optimize and run stencil kernels.
## 1) Kernel Definition ## 1) Kernel Definition
### a) Defining kernels with assignment lists and the `kernel` decorator ### a) Defining kernels with assignment lists and the `kernel` decorator
*pystencils* gets a symbolic formulation of the kernel. This can be either an `Assignment` or a sequence of `Assignment`s that follow a set of restrictions. *pystencils* gets a symbolic formulation of the kernel. This can be either an `Assignment` or a sequence of `Assignment`s that follow a set of restrictions.
Lets first create a kernel that consists of multiple assignments: Lets first create a kernel that consists of multiple assignments:
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
src_arr = np.zeros([20, 30]) src_arr = np.zeros([20, 30])
dst_arr = np.zeros_like(src_arr) dst_arr = np.zeros_like(src_arr)
dst, src = ps.fields(dst=dst_arr, src=src_arr) dst, src = ps.fields(dst=dst_arr, src=src_arr)
``` ```
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
grad_x, grad_y = sp.symbols("grad_x, grad_y") grad_x, grad_y = sp.symbols("grad_x, grad_y")
symbolic_description = [ symbolic_description = [
ps.Assignment(grad_x, (src[1, 0] - src[-1, 0]) / 2), ps.Assignment(grad_x, (src[1, 0] - src[-1, 0]) / 2),
ps.Assignment(grad_y, (src[0, 1] - src[0, -1]) / 2), ps.Assignment(grad_y, (src[0, 1] - src[0, -1]) / 2),
ps.Assignment(dst[0, 0], grad_x + grad_y), ps.Assignment(dst[0, 0], grad_x + grad_y),
] ]
kernel = ps.create_kernel(symbolic_description) kernel = ps.create_kernel(symbolic_description)
symbolic_description symbolic_description
``` ```
%% Output %% Output
$$\left [ grad_{x} \leftarrow \frac{{{src}_{E}}}{2} - \frac{{{src}_{W}}}{2}, \quad grad_{y} \leftarrow \frac{{{src}_{N}}}{2} - \frac{{{src}_{S}}}{2}, \quad {{dst}_{C}} \leftarrow grad_{x} + grad_{y}\right ]$$
$\displaystyle \left[ grad_{x} \leftarrow_{} \frac{{src}_{(1,0)}}{2} - \frac{{src}_{(-1,0)}}{2}, \ grad_{y} \leftarrow_{} \frac{{src}_{(0,1)}}{2} - \frac{{src}_{(0,-1)}}{2}, \ {dst}_{(0,0)} \leftarrow_{} grad_{x} + grad_{y}\right]$
⎡ src_E src_W src_N src_S ⎤ ⎡ src_E src_W src_N src_S ⎤
⎢gradₓ := ───── - ─────, grad_y := ───── - ─────, dst_C := gradₓ + grad_y⎥ ⎢gradₓ := ───── - ─────, grad_y := ───── - ─────, dst_C := gradₓ + grad_y⎥
⎣ 2 2 2 2 ⎦ ⎣ 2 2 2 2 ⎦
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
We created subexpressions, using standard sympy symbols on the left hand side, to split the kernel into multiple assignments. Defining a kernel using a list of `Assignment`s is quite tedious and hard to read. We created subexpressions, using standard sympy symbols on the left hand side, to split the kernel into multiple assignments. Defining a kernel using a list of `Assignment`s is quite tedious and hard to read.
To simplify the formulation of a kernel, *pystencils* offers the `kernel` decorator, that transforms a normal Python function with `@=` assignments into an assignment list that can be passed to `create_kernel`. To simplify the formulation of a kernel, *pystencils* offers the `kernel` decorator, that transforms a normal Python function with `@=` assignments into an assignment list that can be passed to `create_kernel`.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
@ps.kernel @ps.kernel
def symbolic_description_using_function(): def symbolic_description_using_function():
grad_x @= (src[1, 0] - src[-1, 0]) / 2 grad_x @= (src[1, 0] - src[-1, 0]) / 2
grad_y @= (src[0, 1] - src[0, -1]) / 2 grad_y @= (src[0, 1] - src[0, -1]) / 2
dst[0, 0] @= grad_x + grad_y dst[0, 0] @= grad_x + grad_y
symbolic_description_using_function symbolic_description_using_function
``` ```
%% Output %% Output
$$\left [ grad_{x} \leftarrow \frac{{{src}_{E}}}{2} - \frac{{{src}_{W}}}{2}, \quad grad_{y} \leftarrow \frac{{{src}_{N}}}{2} - \frac{{{src}_{S}}}{2}, \quad {{dst}_{C}} \leftarrow grad_{x} + grad_{y}\right ]$$
$\displaystyle \left[ grad_{x} \leftarrow_{} \frac{{src}_{(1,0)}}{2} - \frac{{src}_{(-1,0)}}{2}, \ grad_{y} \leftarrow_{} \frac{{src}_{(0,1)}}{2} - \frac{{src}_{(0,-1)}}{2}, \ {dst}_{(0,0)} \leftarrow_{} grad_{x} + grad_{y}\right]$
⎡ src_E src_W src_N src_S ⎤ ⎡ src_E src_W src_N src_S ⎤
⎢gradₓ := ───── - ─────, grad_y := ───── - ─────, dst_C := gradₓ + grad_y⎥ ⎢gradₓ := ───── - ─────, grad_y := ───── - ─────, dst_C := gradₓ + grad_y⎥
⎣ 2 2 2 2 ⎦ ⎣ 2 2 2 2 ⎦
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
The decorated function can contain any Python code, only the `@=` operator, and the ternary inline `if-else` operator have different meaning. The decorated function can contain any Python code, only the `@=` operator, and the ternary inline `if-else` operator have different meaning.
### b) Ternary 'if' with `Piecewise` ### b) Ternary 'if' with `Piecewise`
The ternary operator maps to `sympy.Piecewise` functions, that can be used to introduce branching into the kernel. Piecewise defined functions must give a value for every input, i.e. there must be a 'otherwise' clause in the end that is indicated by the condition `True`. Piecewise objects are standard sympy terms that can be integrated into bigger expressions: The ternary operator maps to `sympy.Piecewise` functions, that can be used to introduce branching into the kernel. Piecewise defined functions must give a value for every input, i.e. there must be a 'otherwise' clause in the end that is indicated by the condition `True`. Piecewise objects are standard sympy terms that can be integrated into bigger expressions:
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
sp.Piecewise((1.0, src[0,1] > 0), (0.0, True)) + src[1, 0] sp.Piecewise((1.0, src[0,1] > 0), (0.0, True)) + src[1, 0]
``` ```
%% Output %% Output
$${{src}_{E}} + \begin{cases} 1.0 & \text{for}\: {{src}_{N}} > 0 \\0.0 & \text{otherwise} \end{cases}$$
$\displaystyle {src}_{(1,0)} + \begin{cases} 1.0 & \text{for}\: {src}_{(0,1)} > 0 \\0.0 & \text{otherwise} \end{cases}$
⎛⎧1.0 for src_N > 0⎞ ⎛⎧1.0 for src_N > 0⎞
src_E + ⎜⎨ ⎟ src_E + ⎜⎨ ⎟
⎝⎩0.0 otherwise ⎠ ⎝⎩0.0 otherwise ⎠
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
Piecewise objects are created by the `kernel` decorator for ternary if-else statements. Piecewise objects are created by the `kernel` decorator for ternary if-else statements.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
@ps.kernel @ps.kernel
def kernel_with_piecewise(): def kernel_with_piecewise():
grad_x @= (src[1, 0] - src[-1, 0]) / 2 if src[-1, 0] > 0 else 0.0 grad_x @= (src[1, 0] - src[-1, 0]) / 2 if src[-1, 0] > 0 else 0.0
kernel_with_piecewise kernel_with_piecewise
``` ```
%% Output %% Output
$$\left [ grad_{x} \leftarrow \begin{cases} \frac{{{src}_{E}}}{2} - \frac{{{src}_{W}}}{2} & \text{for}\: {{src}_{W}} > 0 \\0.0 & \text{otherwise} \end{cases}\right ]$$
$\displaystyle \left[ grad_{x} \leftarrow_{} \begin{cases} \frac{{src}_{(1,0)}}{2} - \frac{{src}_{(-1,0)}}{2} & \text{for}\: {src}_{(-1,0)} > 0 \\0.0 & \text{otherwise} \end{cases}\right]$
⎡ ⎧src_E src_W ⎤ ⎡ ⎧src_E src_W ⎤
⎢ ⎪───── - ───── for src_W > 0⎥ ⎢ ⎪───── - ───── for src_W > 0⎥
⎢gradₓ := ⎨ 2 2 ⎥ ⎢gradₓ := ⎨ 2 2 ⎥
⎢ ⎪ ⎥ ⎢ ⎪ ⎥
⎣ ⎩ 0.0 otherwise ⎦ ⎣ ⎩ 0.0 otherwise ⎦
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
### c) Assignment level optimizations using `AssignmentCollection` ### c) Assignment level optimizations using `AssignmentCollection`
When the kernels get larger and more complex, it is helpful to organize the list of assignment into a more structured way. The `AssignmentCollection` offers optimizating transformation on a list of assignments. It holds two assignment lists, one for subexpressions and one for the main assignments. Main assignments are typically those that write to an array. When the kernels get larger and more complex, it is helpful to organize the list of assignment into a more structured way. The `AssignmentCollection` offers optimizating transformation on a list of assignments. It holds two assignment lists, one for subexpressions and one for the main assignments. Main assignments are typically those that write to an array.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
@ps.kernel @ps.kernel
def somewhat_longer_dummy_kernel(s): def somewhat_longer_dummy_kernel(s):
s.a @= src[0, 1] + src[-1, 0] s.a @= src[0, 1] + src[-1, 0]
s.b @= 2 * src[1, 0] + src[0, -1] s.b @= 2 * src[1, 0] + src[0, -1]
s.c @= src[0, 1] + 2 * src[1, 0] + src[-1, 0] + src[0, -1] - src[0,0] s.c @= src[0, 1] + 2 * src[1, 0] + src[-1, 0] + src[0, -1] - src[0,0]
dst[0, 0] @= s.a + s.b + s.c dst[0, 0] @= s.a + s.b + s.c
ac = ps.AssignmentCollection(main_assignments=somewhat_longer_dummy_kernel[-1:], ac = ps.AssignmentCollection(main_assignments=somewhat_longer_dummy_kernel[-1:],
subexpressions=somewhat_longer_dummy_kernel[:-1]) subexpressions=somewhat_longer_dummy_kernel[:-1])
ac ac
``` ```
%% Output %% Output
Equation Collection for dst_C AssignmentCollection: dst_C, <- f(src_C, src_W, src_S, src_N, src_E)
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
ac.operation_count ac.operation_count
``` ```
%% Output %% Output
{'adds': 8, 'muls': 2, 'divs': 0} {'adds': 8,
'muls': 2,
'divs': 0,
'sqrts': 0,
'fast_sqrts': 0,
'fast_inv_sqrts': 0,
'fast_div': 0}
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
The `pystencils.simp` submodule offers several functions to optimize a collection of assignments. The `pystencils.simp` submodule offers several functions to optimize a collection of assignments.
It also offers functionality to group optimization into strategies and evaluate them. It also offers functionality to group optimization into strategies and evaluate them.
In this example we reduce the number of operations by reusing existing subexpressions to get rid of two unnecessary floating point additions. For more information about assignment collections and simplifications see the [demo notebook](demo_assignment_collection.ipynb). In this example we reduce the number of operations by reusing existing subexpressions to get rid of two unnecessary floating point additions. For more information about assignment collections and simplifications see the [demo notebook](demo_assignment_collection.ipynb).
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
opt_ac = ps.simp.subexpression_substitution_in_existing_subexpressions(ac) opt_ac = ps.simp.subexpression_substitution_in_existing_subexpressions(ac)
opt_ac opt_ac
``` ```
%% Output %% Output
Equation Collection for dst_C AssignmentCollection: dst_C, <- f(src_C, src_W, src_S, src_N, src_E)
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
opt_ac.operation_count opt_ac.operation_count
``` ```
%% Output %% Output
{'adds': 6, 'muls': 1, 'divs': 0} {'adds': 6,
'muls': 1,
'divs': 0,
'sqrts': 0,
'fast_sqrts': 0,
'fast_inv_sqrts': 0,
'fast_div': 0}
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
### d) Ghost layers and iteration region ### d) Ghost layers and iteration region
When creating a kernel with neighbor accesses, *pystencils* automatically restricts the iteration region, such that all accesses are safe. When creating a kernel with neighbor accesses, *pystencils* automatically restricts the iteration region, such that all accesses are safe.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
kernel = ps.create_kernel(ps.Assignment(dst[0,0], src[2, 0] + src[-1, 0])) kernel = ps.create_kernel(ps.Assignment(dst[0,0], src[2, 0] + src[-1, 0]))
ps.show_code(kernel) ps.show_code(kernel)
``` ```
%% Output %% Output
FUNC_PREFIX void kernel(double * RESTRICT fd_dst, double * RESTRICT const fd_src)
{
for (int ctr_0 = 2; ctr_0 < 18; ctr_0 += 1)
{
double * RESTRICT fd_dst_C = 30*ctr_0 + fd_dst;
double * RESTRICT const fd_src_2E = 30*ctr_0 + fd_src + 60;
double * RESTRICT const fd_src_W = 30*ctr_0 + fd_src - 30;
for (int ctr_1 = 2; ctr_1 < 28; ctr_1 += 1)
{
fd_dst_C[ctr_1] = fd_src_2E[ctr_1] + fd_src_W[ctr_1];
}
}
}
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
When no additional ghost layer information is given, *pystencils* looks at all neighboring field accesses and introduces the required number of ghost layers **for all directions**. In the example above the largest neighbor accesses was ``src[2, 0]``, so theoretically we would need 2 ghost layers only the the end of the x coordinate. When no additional ghost layer information is given, *pystencils* looks at all neighboring field accesses and introduces the required number of ghost layers **for all directions**. In the example above the largest neighbor accesses was ``src[2, 0]``, so theoretically we would need 2 ghost layers only the the end of the x coordinate.
By default *pystencils* introduces 2 ghost layers at all borders of the domain. The next cell shows how to change this behavior. Be careful with manual ghost layer specification, wrong values may lead to SEGFAULTs. By default *pystencils* introduces 2 ghost layers at all borders of the domain. The next cell shows how to change this behavior. Be careful with manual ghost layer specification, wrong values may lead to SEGFAULTs.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
gl_spec = [(0, 2), # 0 ghost layers at the left, 2 at the right border gl_spec = [(0, 2), # 0 ghost layers at the left, 2 at the right border
(1, 0)] # 1 ghost layer at the lower y, one at the upper y coordinate (1, 0)] # 1 ghost layer at the lower y, one at the upper y coordinate
kernel = ps.create_kernel(ps.Assignment(dst[0,0], src[2, 0] + src[-1, 0]), ghost_layers=gl_spec) kernel = ps.create_kernel(ps.Assignment(dst[0,0], src[2, 0] + src[-1, 0]), ghost_layers=gl_spec)
ps.show_code(kernel) ps.show_code(kernel)
``` ```
%% Output %% Output
FUNC_PREFIX void kernel(double * RESTRICT fd_dst, double * RESTRICT const fd_src)
{
for (int ctr_0 = 0; ctr_0 < 18; ctr_0 += 1)
{
double * RESTRICT fd_dst_C = 30*ctr_0 + fd_dst;
double * RESTRICT const fd_src_2E = 30*ctr_0 + fd_src + 60;
double * RESTRICT const fd_src_W = 30*ctr_0 + fd_src - 30;
for (int ctr_1 = 1; ctr_1 < 30; ctr_1 += 1)
{
fd_dst_C[ctr_1] = fd_src_2E[ctr_1] + fd_src_W[ctr_1];
}
}
}
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
## 2 ) Restrictions ## 2 ) Restrictions
### a) Independence Restriction ### a) Independence Restriction
*pystencils* only works for kernels where each array element can be updated independently from all other elements. This restriction ensures that the kernels can be easily parallelized and also be run on the GPU. Trying to define kernels where the results depends on the iteration order, leads to a ValueError. *pystencils* only works for kernels where each array element can be updated independently from all other elements. This restriction ensures that the kernels can be easily parallelized and also be run on the GPU. Trying to define kernels where the results depends on the iteration order, leads to a ValueError.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
invalid_description = [ invalid_description = [
ps.Assignment(dst[1, 0], src[1, 0] + src[-1, 0]), ps.Assignment(dst[1, 0], src[1, 0] + src[-1, 0]),
ps.Assignment(dst[0, 0], src[1, 0] - src[-1, 0]), ps.Assignment(dst[0, 0], src[1, 0] - src[-1, 0]),
] ]
try: try:
invalid_kernel = ps.create_kernel(invalid_description) invalid_kernel = ps.create_kernel(invalid_description)
assert False, "Should never be executed" assert False, "Should never be executed"
except ValueError as e: except ValueError as e:
print(e) print(e)
``` ```
%% Output %% Output
Field dst is written at two different locations Field dst is written at two different locations
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
The independence restriction makes sure that the kernel can be safely parallelized by checking the following conditions: If a field is modified inside the kernel, it may only be modified at a single spatial position. In that case the field may also only be read at this position. Fields that are not modified may be read at multiple neighboring positions. The independence restriction makes sure that the kernel can be safely parallelized by checking the following conditions: If a field is modified inside the kernel, it may only be modified at a single spatial position. In that case the field may also only be read at this position. Fields that are not modified may be read at multiple neighboring positions.
Specifically, this rule allows for in-place updates that don't access neighbors. Specifically, this rule allows for in-place updates that don't access neighbors.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
valid_kernel = ps.create_kernel(ps.Assignment(src[0,0], 2*src[0,0] + 42)) valid_kernel = ps.create_kernel(ps.Assignment(src[0,0], 2*src[0,0] + 42))
``` ```
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
If a field stores multiple values per cell, as in the next example, this restriction only applies for accesses with the same index. If a field stores multiple values per cell, as in the next example, this restriction only applies for accesses with the same index.
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
v = ps.fields("v(2): double[2D]") v = ps.fields("v(2): double[2D]")
valid_kernel = ps.create_kernel([ps.Assignment(v[0,0](1), 2*v[0,0](1) + 42), valid_kernel = ps.create_kernel([ps.Assignment(v[0,0](1), 2*v[0,0](1) + 42),
ps.Assignment(v[0,1](0), 2*v[1,0](0) + 42)]) ps.Assignment(v[0,1](0), 2*v[0,1](0) + 42)])
``` ```
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
### b) Static Single Assignment Form ### b) Static Single Assignment Form
All assignments that don't write to a field must be in SSA form All assignments that don't write to a field must be in SSA form
1. Each sympy symbol may only occur once as a left-hand-side (fields can be written multiple times) 1. Each sympy symbol may only occur once as a left-hand-side (fields can be written multiple times)
2. A symbol has to be defined before it is used. If it is never defined it is introduced as function parameter 2. A symbol has to be defined before it is used. If it is never defined it is introduced as function parameter
The next cell demonstrates the first SSA restriction: The next cell demonstrates the first SSA restriction:
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
@ps.kernel @ps.kernel
def not_allowed(): def not_allowed():
a, b = sp.symbols("a b") a, b = sp.symbols("a b")
a @= src[0, 0] a @= src[0, 0]
b @= a + 3 b @= a + 3
a @= src[-1, 0] a @= src[-1, 0]
dst[0, 0] @= a + b dst[0, 0] @= a + b
try: try:
ps.create_kernel(not_allowed) ps.create_kernel(not_allowed)
assert False assert False
except ValueError as e: except ValueError as e:
print(e) print(e)
``` ```
%% Output %% Output
Assignments not in SSA form, multiple assignments to a Assignments not in SSA form, multiple assignments to a
%% Cell type:markdown id: tags: %% Cell type:markdown id: tags:
However, for right hand sides that are Field.Accesses this is allowed: Also it is not allowed to write a field at the same location
%% Cell type:code id: tags: %% Cell type:code id: tags:
``` python ``` python
@ps.kernel @ps.kernel
def allowed(): def not_allowed():
dst[0, 0] @= src[0, 1] + src[1, 0] dst[0, 0] @= src[0, 1] + src[1, 0]
dst[0, 0] @= 2 * dst[0, 0] dst[0, 0] @= 2 * dst[0, 0]
ps.create_kernel(allowed)
try:
ps.create_kernel(not_allowed)
assert False
except ValueError as e:
print(e)
``` ```
%% Output %% Output
KernelFunction kernel([<double * RESTRICT fd_dst>, <double * RESTRICT const fd_src>]) Field dst is written twice at the same location
%% Cell type:markdown id: tags:
This situation should be resolved by introducing temporary variables
%% Cell type:code id: tags:
``` python
tmp_var = sp.Symbol("a")
@ps.kernel
def allowed():
tmp_var @= src[0, 1] + src[1, 0]
dst[0, 0] @= 2 * tmp_var
ast = ps.create_kernel(allowed)
ps.show_code(ast)
```
%% Output
......