Ralf Gommers rgommers

## blas_lapack_meson_notes.md

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                rgommers
                / blas_lapack_meson_notes.md
            
            
              Last active
              October 13, 2023 18:08
            
              
                Notes on BLAS/LAPACK library details and conventions
              
          
OpenBLAS can be built with NOFORTRAN, in that case it's CBLAS + f2c'd LAPACK
OpenBLAS can be built without LAPACK support, Arch Linux currently does this
(see scipy/scipy#17465)
OpenBLAS library can be renamed with an option in its Makefile
Build options:

conda-forge: https://github.com/conda-forge/openblas-feedstock/blob/49ca08fc9d1ff220804aa9b894b9a6fe5db45057/recipe/conda_build_config.yaml
Spack: https://github.com/spack/spack/blob/develop/var/spack/repos/builtin/packages/openblas/package.py#L52
vendored into NumPy/SciPy wheels: https://github.com/MacPython/openblas-libs/blob/master/tools/build_openblas.sh#L53


Conda-forge library names and pkg-config output

  
## spack-scipy-oneapi-build.log
==> py-scipy: Executing phase: 'install'
==> [2022-10-13-12:45:50.325962] '/home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/python-3.9.13-mn4ovd4dkbyg62orinqikrnvaxq6r5tf/bin/python3.9' '-m' 'pip' '-vvv' '--no-input' '--no-cache-dir' '--disable-pip-version-check' 'install' '--no-deps' '--ignore-installed' '--no-build-isolation' '--no-warn-script-location' '--no-index' '--prefix=/home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/py-scipy-1.8.1-ww5y2ruckqwn44vypo4e7yl3yzmxiwsh' '.'
Using pip 22.2.2 from /home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/py-pip-22.2.2-42kbxmhhxyigyp2224oarapc3o5xxpga/lib/python3.9/site-packages/pip (python 3.9)
Non-user install due to --prefix or --target option
Ignoring indexes: https://pypi.org/simple
Created temporary directory: /tmp/pip-ephem-wheel-cache-vju1m2iu
Created temporary directory: /tmp/pip-build-tracker-h5l54x96
Initialized build tracking at

## jax_numpy_random_apis.py
"""
Implement `jax.random` APIs with NumPy, and `numpy.random` APIs with JAX.
The purpose of this is to be able to compare APIs more easily, and clarify
where they are and aren't similar.
"""

import secrets
import multiprocessing

import numpy as np

## install_array_libraries.txt
To create an environment with the most recent version of all mainstream Python array/tensor libraries installed:

  conda create -n many-libs python=3.7
  conda activate many-libs
  conda install cudatoolkit=10.2
  pip install numpy torch jax jaxlib tensorflow mxnet cupy-cuda102 dask toolz sparse

Conda doesn't manage to find a winning combination here; pip has a hard time
too and probably not all constraints are satisfied, but nothing crashes
and basic tests work as they are supposed to.

## gist:032f0d256c3b11e76d64544b9185f06e
==> 2018-03-06 17:49:03 <==
# cmd: /home/rgommers/anaconda3/bin/conda install pytorch torchvision -c pytorch --no-deps
+defaults::_ipyw_jlab_nb_ext_conf-0.1.0-py36he11e457_0
+defaults::alabaster-0.7.10-py36h306e16b_0
+defaults::anaconda-5.1.0-py36_2
+defaults::anaconda-client-1.6.9-py36_0
+defaults::anaconda-navigator-1.7.0-py36_0
+defaults::anaconda-project-0.8.2-py36h44fb852_0
+defaults::asn1crypto-0.24.0-py36_0
+defaults::astroid-1.6.1-py36_0

## scipy_numpy_parallel_build_timings
# scipy
time python setup.py build

real    16m52.989s
user    8m37.808s
sys     0m10.156s

time python setup.py build -j2

real    13m27.485s

## autograd_specialfuncs.py
import numpy as np
import matplotlib.pyplot as plt
from autograd import grad
import autograd.scipy.special as special


plt.style.use('ggplot')


x = np.linspace(-10, 10, num=1000)

## keybase.md

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                rgommers
                / keybase.md
            
            
              Created
              June 25, 2017 03:42
            
          
    Keybase proof

I hereby claim:

I am rgommers on github.
I am rgommers (https://keybase.io/rgommers) on keybase.
I have a public key whose fingerprint is BFE9 5DF1 9868 9E30 AFF4  A1CD EDBE A378 BF1A 5EBD

To claim this, I am signing this object:

  
## gist:bca29ea209e7714848c5884b88d338bb
Output of ``scipy.test('full', timer=250)`` against numpy master with scipy version ``1.0.0.dev0+d88eb06``.


    [success] 1.90% test_kdeoth.test_kde_2d: 10.5936s
    [success] 1.89% test_mpmath.TestSystematic.test_betainc: 10.5385s
    [success] 1.86% test_linsolve.TestSpsolveTriangular.test_random: 10.3594s
    [success] 1.74% test_cdflib.TestCDFlib.test_chndtrix: 9.7304s
    [success] 1.67% test_mpmath.TestSystematic.test_bessely_complex: 9.3187s
    [success] 1.58% test_circular_imports.test_importing_submodules: 8.8049s
    [success] 1.51% test_basic.TestSolve.test_all_type_size_routine_combinations: 8.4452s

## gist:b06981f89fb67eb6c1e643cab565419d
With numpy 1.12.1 and scipy master.


$ python -c "import skimage; skimage.test()"
.................................................................................................................................................................................................................................../home/rgommers/Code/scikit-image/skimage/transform/_warps.py:84: UserWarning: The default mode, 'constant', will be changed to 'reflect' in skimage 0.15.
  warn("The default mode, 'constant', will be changed to 'reflect' in "
......../home/rgommers/Code/scikit-image/skimage/feature/corner.py:373: RuntimeWarning: divide by zero encountered in divide
  return (2.0 / np.pi) * np.arctan((l2 + l1) / (l2 - l1))
/home/rgommers/Code/scikit-image/skimage/feature/corner.py:373: RuntimeWarning: invalid value encountered in divide
  return (2.0 / np.pi) * np.arctan((l2 + l1) / (l2 - l1))
	==> py-scipy: Executing phase: 'install'
	==> [2022-10-13-12:45:50.325962] '/home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/python-3.9.13-mn4ovd4dkbyg62orinqikrnvaxq6r5tf/bin/python3.9' '-m' 'pip' '-vvv' '--no-input' '--no-cache-dir' '--disable-pip-version-check' 'install' '--no-deps' '--ignore-installed' '--no-build-isolation' '--no-warn-script-location' '--no-index' '--prefix=/home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/py-scipy-1.8.1-ww5y2ruckqwn44vypo4e7yl3yzmxiwsh' '.'
	Using pip 22.2.2 from /home/rgommers/code/spack/opt/spack/linux-endeavourosrolling-skylake_avx512/oneapi-2022.2.0/py-pip-22.2.2-42kbxmhhxyigyp2224oarapc3o5xxpga/lib/python3.9/site-packages/pip (python 3.9)
	Non-user install due to --prefix or --target option
	Ignoring indexes: https://pypi.org/simple
	Created temporary directory: /tmp/pip-ephem-wheel-cache-vju1m2iu
	Created temporary directory: /tmp/pip-build-tracker-h5l54x96
	Initialized build tracking at
	"""
	Implement `jax.random` APIs with NumPy, and `numpy.random` APIs with JAX.
	The purpose of this is to be able to compare APIs more easily, and clarify
	where they are and aren't similar.
	"""

	import secrets
	import multiprocessing

	import numpy as np
	To create an environment with the most recent version of all mainstream Python array/tensor libraries installed:

	conda create -n many-libs python=3.7
	conda activate many-libs
	conda install cudatoolkit=10.2
	pip install numpy torch jax jaxlib tensorflow mxnet cupy-cuda102 dask toolz sparse

	Conda doesn't manage to find a winning combination here; pip has a hard time
	too and probably not all constraints are satisfied, but nothing crashes
	and basic tests work as they are supposed to.
	==> 2018-03-06 17:49:03 <==
	# cmd: /home/rgommers/anaconda3/bin/conda install pytorch torchvision -c pytorch --no-deps
	+defaults::_ipyw_jlab_nb_ext_conf-0.1.0-py36he11e457_0
	+defaults::alabaster-0.7.10-py36h306e16b_0
	+defaults::anaconda-5.1.0-py36_2
	+defaults::anaconda-client-1.6.9-py36_0
	+defaults::anaconda-navigator-1.7.0-py36_0
	+defaults::anaconda-project-0.8.2-py36h44fb852_0
	+defaults::asn1crypto-0.24.0-py36_0
	+defaults::astroid-1.6.1-py36_0
	# scipy
	time python setup.py build

	real 16m52.989s
	user 8m37.808s
	sys 0m10.156s

	time python setup.py build -j2

	real 13m27.485s
	import numpy as np
	import matplotlib.pyplot as plt
	from autograd import grad
	import autograd.scipy.special as special


	plt.style.use('ggplot')


	x = np.linspace(-10, 10, num=1000)
	Output of ``scipy.test('full', timer=250)`` against numpy master with scipy version ``1.0.0.dev0+d88eb06``.


	[success] 1.90% test_kdeoth.test_kde_2d: 10.5936s
	[success] 1.89% test_mpmath.TestSystematic.test_betainc: 10.5385s
	[success] 1.86% test_linsolve.TestSpsolveTriangular.test_random: 10.3594s
	[success] 1.74% test_cdflib.TestCDFlib.test_chndtrix: 9.7304s
	[success] 1.67% test_mpmath.TestSystematic.test_bessely_complex: 9.3187s
	[success] 1.58% test_circular_imports.test_importing_submodules: 8.8049s
	[success] 1.51% test_basic.TestSolve.test_all_type_size_routine_combinations: 8.4452s
	With numpy 1.12.1 and scipy master.


	$ python -c "import skimage; skimage.test()"
	.................................................................................................................................................................................................................................../home/rgommers/Code/scikit-image/skimage/transform/_warps.py:84: UserWarning: The default mode, 'constant', will be changed to 'reflect' in skimage 0.15.
	warn("The default mode, 'constant', will be changed to 'reflect' in "
	......../home/rgommers/Code/scikit-image/skimage/feature/corner.py:373: RuntimeWarning: divide by zero encountered in divide
	return (2.0 / np.pi) * np.arctan((l2 + l1) / (l2 - l1))
	/home/rgommers/Code/scikit-image/skimage/feature/corner.py:373: RuntimeWarning: invalid value encountered in divide
	return (2.0 / np.pi) * np.arctan((l2 + l1) / (l2 - l1))