Merge changes to master (#1)

* Try to merge * fixed merge --------- Co-authored-by: chaithyagr <[email protected]>
flatironinstitute · Nov 27, 2023 · fe8abc5 · fe8abc5
1 parent 7273f68
commit fe8abc5
Show file tree

Hide file tree

Showing 112 changed files with 3,129 additions and 3,873 deletions.
diff --git a/.github/workflows/python_build_win.ps1 b/.github/workflows/python_build_win.ps1
@@ -41,6 +41,8 @@ Copy-Item -Path C:\msys64\mingw64\bin\libgomp-*.dll -Destination ([IO.Path]::Com
 Copy-Item -Path C:\msys64\mingw64\bin\libwinpthread-*.dll -Destination ([IO.Path]::Combine($unpacked_wheel, 'finufft'))
 Copy-Item -Path C:\msys64\mingw64\bin\libfftw3-*.dll -Destination ([IO.Path]::Combine($unpacked_wheel, 'finufft'))
 Copy-Item -Path C:\msys64\mingw64\bin\libfftw3f-*.dll -Destination ([IO.Path]::Combine($unpacked_wheel, 'finufft'))
+Copy-Item -Path C:\msys64\mingw64\bin\libfftw3_omp-*.dll -Destination ([IO.Path]::Combine($unpacked_wheel, 'finufft'))
+Copy-Item -Path C:\msys64\mingw64\bin\libfftw3f_omp-*.dll -Destination ([IO.Path]::Combine($unpacked_wheel, 'finufft'))
 New-Item -Path .\wheelhouse -ItemType Directory -Force
 wheel.exe pack $unpacked_wheel -d .\wheelhouse
 if (-not $?) {throw "Failed pack wheel"}

diff --git a/.github/workflows/python_test_win.ps1 b/.github/workflows/python_test_win.ps1
@@ -1,6 +1,9 @@
-python -m pip install finufft -f .\wheelhouse\
+python -m pip install --pre finufft -f .\wheelhouse\
 if (-not $?) {throw "Failed to pip install finufft"}
 python python/finufft/test/run_accuracy_tests.py
 if (-not $?) {throw "Tests failed"}
 python python/finufft/examples/simple1d1.py
 if (-not $?) {throw "Simple1d1 test failed"}
+python -m pip install pytest
+python -m pytest python/finufft/test
+if (-not $?) {throw "Pytest suite failed"}
diff --git a/.github/workflows/python_wheel.yml b/.github/workflows/python_wheel.yml
@@ -34,7 +34,7 @@ jobs:
       uses: actions/upload-artifact@v2
       with:
         name: linux-wheels
-        path: python/wheelhouse/finufft*manylinux*.whl
+        path: python/finufft/wheelhouse/finufft*manylinux*.whl
 
   MacOS:
     runs-on: macos-latest
@@ -60,19 +60,6 @@ jobs:
         # hack to make libquadmath link statically
         sudo rm /usr/local/opt/gcc@11/lib/gcc/11/libquadmath.*dylib
 
-    - name: Cache macOS Python binaries
-      id: cache-python
-      uses: actions/cache@v3
-      with:
-        path: |
-          /Library/Frameworks/Python.framework/Versions/3.6
-          /Library/Frameworks/Python.framework/Versions/3.7
-          /Library/Frameworks/Python.framework/Versions/3.8
-          /Library/Frameworks/Python.framework/Versions/3.9
-          /Library/Frameworks/Python.framework/Versions/3.10
-          /Library/Frameworks/Python.framework/Versions/3.11
-        key: macos-python-3.6.8-macosx10.9-python3.7.9-macosx10.9-python3.8.3-macosx10.9-python3.9.7-macos11-python3.10.1-macos11-python3.11.0-macos11
-
     # Download and install Python instead of using the setup_python
     # as the python interpreters in the Github machines
     # were compiled in 10.14, the wheels built with them
@@ -138,30 +125,42 @@ jobs:
         PYTHON_BIN=/Library/Frameworks/Python.framework/Versions/3.11/bin/
         $PYTHON_BIN/python3 -m pip install delocate
         ls wheelhouse/finufft*.whl | xargs -n1 $PYTHON_BIN/delocate-wheel -w fixed_wheel/
-        /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 examples/simple1d1.py
-        /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.6/bin/python3 -m pytest test
+        /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 examples/simple1d1.py
-        /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.7/bin/python3 -m pytest test
+        /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 examples/simple1d1.py
-        /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.8/bin/python3 -m pytest test
+        /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 examples/simple1d1.py
-        /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.9/bin/python3 -m pytest test
+        /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 examples/simple1d1.py
-        /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 -m pip install finufft -f fixed_wheel/
+        /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.10/bin/python3 -m pytest test
+        /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 -m pip install --pre finufft -f fixed_wheel/
         /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 test/run_accuracy_tests.py
         /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 examples/simple1d1.py
+        /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 -m pip install pytest
+        /Library/Frameworks/Python.framework/Versions/3.11/bin/python3 -m pytest test
 
     - name: Upload wheels
       uses: actions/upload-artifact@v2
       with:
         name: macos-wheels
-        path: python/fixed_wheel/*.whl
+        path: python/finufft/fixed_wheel/*.whl
 
   Windows:
     runs-on: windows-latest

diff --git a/.readthedocs.yml → .readthedocs.yaml b/.readthedocs.yml → .readthedocs.yaml
@@ -1,12 +1,16 @@
 # .readthedocs.yml
 # "Read the Docs" doc-hosting website configuration file
 # See https://docs.readthedocs.io/en/stable/config-file/v2.html for details
-#
-# Barnett 10/5/20.
 
 # Required (for this file format)
 version: 2
 
+# Set the OS, Python version and other tools you might need
+build:
+  os: ubuntu-22.04
+  tools:
+    python: "3.11"
+
 # Build all formats
 formats: all
 
@@ -16,6 +20,5 @@ sphinx:
 
 # Optionally set the version of Python and requirements required to build your docs
 python:
-  version: 3.7
   install:
     - requirements: docs/requirements.txt
diff --git a/CHANGELOG b/CHANGELOG
@@ -1,7 +1,12 @@
 List of features / changes made / release notes, in reverse chronological order.
 If not stated, FINUFFT is assumed (up to v 1.3 CUFINUFFT indicated separately).
 
-V 2.2.0 (not yet released... planning summer 2023):
+V 2.3.0 (expected Dec 2023)
+
+* finally thread-safety for all fftw cases, kill FFTW_PLAN_SAFE (PR 354)
+* python interface edge cases (singleton dims, #395). Awaiting #367
+
+V 2.2.0 (08/30/23) (untagged; aka 2.2.0.dev0, a name used to bump python pkg)
 
 * MERGE OF CUFINUFFT (GPU CODE) INTO FINUFFT SOURCE TREE:
   - combined cmake build system via FINUFFT_USE_CUDA flag
@@ -13,7 +18,7 @@ V 2.2.0 (not yet released... planning summer 2023):
   - cufinufft repo will be obsoleted.
   - coding leads on this: Robert Blackwell and Joakim Anden
 * cmake build structure (thanks: Wenda Zhou, Marco Barbone, Libin Lu)
-  - Note: the plan is to make the makefiles/make.inc.* obsolete by 2023 end.
+  - Note: the plan is to make the makefiles/make.inc.* obsolete.
 * interp (for type 2) accel by up to 60% in high-acc 2D/3D, by FMA/SIMD-friendly
   rearranging of loops, by Martin Reinecke, PR #292.
 * remove inv array in binsort; speeds up multithreaded case by up to 50%
@@ -25,7 +30,7 @@ V 2.2.0 (not yet released... planning summer 2023):
   remains public, as should be (PR #233). Allows plan to have C++ constructs.
 * fixed single-thread (OMP=OFF) build which failed due to fftw_defs.h/defs.h
 
-CUFINUFFT v 1.3 (06/10/23)
+CUFINUFFT v 1.3 (06/10/23) (Final legacy release outside FINUFFT repo)
 
 * Move second half of onedim_fseries_kernel() to GPU (with a simple heuristic
   basing on nf1 to switch between the CPU and the GPU version).

diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.19)
 
-project(finufft VERSION 2.1.0 LANGUAGES C CXX)
+project(finufft VERSION 2.2.0 LANGUAGES C CXX)
 
 set(GNU_LIKE_FRONTENDS AppleClang Clang GNU)
 if(CMAKE_CXX_COMPILER_ID IN_LIST GNU_LIKE_FRONTENDS)
@@ -11,7 +11,12 @@ endif()
 
 include(CTest)
 
-set(FINUFFT_ARCH_FLAGS "-march=native" CACHE STRING "Compiler flags for specifying target architecture.")
+if(CMAKE_SYSTEM_PROCESSOR MATCHES "ppc|ppc64|powerpc|powerpc64" OR (APPLE AND CMAKE_OSX_ARCHITECTURES MATCHES "ppc|ppc64"))
+    # PowerPC arch does not have -march flag.
+    set(FINUFFT_ARCH_FLAGS "-mtune=native" CACHE STRING "Compiler flags for specifying target architecture.")
+else()
+    set(FINUFFT_ARCH_FLAGS "-march=native" CACHE STRING "Compiler flags for specifying target architecture.")
+endif()
 set(FINUFFT_FFTW_SUFFIX "OpenMP" CACHE STRING "Suffix for FFTW libraries (e.g. OpenMP, Threads etc.)")
 set(FINUFFT_FFTW_LIBRARIES "DEFAULT" CACHE STRING "Specify a custom FFTW library")
 

diff --git a/CMakePresets.json b/CMakePresets.json
@@ -55,6 +55,20 @@
                 "FINUFFT_USE_OPENMP": "OFF"
             }
         },
+        {
+            "name": "icx",
+            "binaryDir": "build/icx",
+            "displayName": "Intel Compiler (llvm)",
+            "description": "Build with Intel Compiler",
+            "generator": "Ninja Multi-Config",
+            "cacheVariables": {
+                "CMAKE_C_COMPILER": "icx",
+                "CMAKE_CXX_COMPILER": "icpx",
+                "CMAKE_Fortran_COMPILER": "ifx",
+                "FINUFFT_ARCH_FLAGS": "-xHost",
+                "CMAKE_CXX_FLAGS": "-fp-model=strict"
+            }
+        },
         {
             "name": "icc",
             "binaryDir": "build/icc",
@@ -109,6 +123,11 @@
             "configurePreset": "icc",
             "configuration": "RelWithDebInfo"
         },
+        {
+            "name": "icx",
+            "configurePreset": "icx",
+            "configuration": "RelWithDebInfo"
+        },
         {
             "name": "matlab",
             "configurePreset": "matlab",

diff --git a/Jenkinsfile b/Jenkinsfile
@@ -9,30 +9,53 @@ pipeline {
     stage('main') {
       agent {
          dockerfile {
-            filename 'tools/cufinufft/docker/cuda11.0/Dockerfile-x86_64'
+            filename 'tools/cufinufft/docker/cuda11.2/Dockerfile-x86_64'
             args '--gpus 2'
+            label 'v100'
          }
       }
       environment {
-    HOME = "$WORKSPACE/build"
+    HOME = "$WORKSPACE"
     PYBIN = "/opt/python/cp38-cp38/bin"
+    LIBRARY_PATH = "$WORKSPACE/build"
+    LD_LIBRARY_PATH = "$WORKSPACE/build"
       }
       steps {
     sh '''#!/bin/bash -ex
       nvidia-smi
     '''
     sh '''#!/bin/bash -ex
-      cp -r /io/build/test/cuda cuda_tests
-      cd cuda_tests
+      echo $HOME
+    '''
+    sh '''#!/bin/bash -ex
+        # v100 cuda arch
+        cuda_arch="70"
+
+        cmake -B build . -DFINUFFT_USE_CUDA=ON \
+                         -DFINUFFT_USE_CPU=OFF \
+                         -DFINUFFT_BUILD_TESTS=ON \
+                         -DCMAKE_CUDA_ARCHITECTURES="$cuda_arch" \
+                         -DBUILD_TESTING=ON
+        cd build
+        make -j4
+    '''
+    sh '''#!/bin/bash -ex
+      cd build/test/cuda
       ctest --output-on-failure
     '''
     sh '${PYBIN}/python3 -m venv $HOME'
     sh '''#!/bin/bash -ex
       source $HOME/bin/activate
       python3 -m pip install --upgrade pip
-      LIBRARY_PATH=/io/build python3 -m pip install -e python/cufinufft
+      python3 -m pip install --upgrade pycuda cupy-cuda112 numba
+      python3 -m pip install torch==1.7.1+cu110 -f https://download.pytorch.org/whl/torch_stable.html
+      python3 -m pip install -e python/cufinufft
       python3 -m pip install pytest
-      python3 -m pytest
+      python -c "from numba import cuda; cuda.cudadrv.libs.test()"
+      python3 -m pytest --framework=pycuda python/cufinufft
+      python3 -m pytest --framework=numba python/cufinufft
+      python3 -m pytest --framework=cupy python/cufinufft
+      python3 -m pytest --framework=torch python/cufinufft
     '''
       }
     }

diff --git a/README.md b/README.md
@@ -3,20 +3,20 @@
 
 Principal author **Alex H. Barnett**,
 main co-developers Jeremy F. Magland,
-Ludvig af Klinteberg, Yu-hsuan "Melody" Shih, Andrea Malleo, Libin Lu,
-and Joakim Andén;
+Ludvig af Klinteberg, Yu-hsuan "Melody" Shih, Libin Lu,
+Joakim Andén, and Robert Blackwell;
 see `docs/ackn.rst` for full list of contributors.
 
 <img align="right" src="docs/logo.png" width="350">
 
 <img align="right" src="docs/spreadpic.png" width="400"/>
 
-This is a lightweight CPU library to compute the three standard types of nonuniform FFT to a specified precision, in one, two, or three dimensions. It is written in C++ with interfaces to C, Fortran, MATLAB/octave, Python, and (in a separate [repository](https://github.com/ludvigak/FINUFFT.jl)) Julia. It now contains the GPU CUDA library cuFINUFFT.
+This is a lightweight CPU library to compute the three standard types of nonuniform FFT to a specified precision, in one, two, or three dimensions. It is written in C++ with interfaces to C, Fortran, MATLAB/octave, Python, and (in a separate [repository](https://github.com/ludvigak/FINUFFT.jl)) Julia. It now also integrates the GPU CUDA library cuFINUFFT (which currently does all but type 3).
 
 Please see the [online documentation](http://finufft.readthedocs.io/en/latest/index.html) which can also be downloaded as a [PDF manual](https://finufft.readthedocs.io/_/downloads/en/latest/pdf/).
-You will also want to see example codes in the directories
-`examples`, `test`, `fortran`, `matlab/test`, and `python/test`.
-If you cannot compile, or `pip install`, try our rather outdated [precompiled binaries](http://users.flatironinstitute.org/~ahb/codes/finufft-binaries).
+You will also want to see CPU example codes in the directories `examples`, `test`, `fortran`, `matlab/test`, `matlab/examples`, `python/finufft/test`, etc, and GPU examples in `examples/cuda`, `test/cuda`, etc
+
+If you cannot build via cMake, try the old makefile. Python users try `pip install finufft`. See the docs for details. See our GitHub Issues for tips.
 
 If you prefer to read text files, the source to generate the above documentation is in human-readable (mostly .rst) files as follows:
 
@@ -29,22 +29,22 @@ If you prefer to read text files, the source to generate the above documentation
 - `docs/c_gpu.rst`   : documentation of C++/C function API for GPU library
 - `docs/opts.rst`    : optional parameters
 - `docs/error.rst`   : error codes
-- `docs/trouble.rst` : troubleshooting
+- `docs/trouble.rst` : troubleshooting advice
 - `docs/tut.rst` and `docs/tutorial/*` : tutorial application examples
 - `docs/fortran.rst` : usage examples from Fortran, documentation of interface
 - `docs/matlab.rst` and `docs/matlabhelp.raw` : using the MATLAB/Octave interface
 - `docs/python.rst` and `python/*/_interfaces.py` : using the Python interface
 - `docs/python_gpu.rst` : Python interface to GPU library
-- `docs/julia.rst`   : using the Julia interface
+- `docs/julia.rst`   : options for Julia users
 - `docs/devnotes.rst`: notes/guide for developers
 - `docs/related.rst` : other recommended NUFFT packages
-- `docs/users.rst`   : users of FINUFFT and dependent packages
+- `docs/users.rst`   : some known users of FINUFFT, dependent packages
 - `docs/ackn.rst`    : authors and acknowledgments
-- `docs/refs.rst`    : journal article references (ours and others)
+- `docs/refs.rst`    : journal article references (both ours and others)
 
 
-If you find (cu)FINUFFT useful in your work, please cite this repository and
-the following. For FINUFFT (CPU library):
+If you find (cu)FINUFFT useful in your work, please star this repository and
+cite it and the following. For FINUFFT (CPU library):
 
 A parallel non-uniform fast Fourier transform library based on an ``exponential of semicircle'' kernel.
 A. H. Barnett, J. F. Magland, and L. af Klinteberg.

diff --git a/TODO → devel/TODO b/TODO → devel/TODO