Skip to content

Instantly share code, notes, and snippets.

@huyng
Last active March 14, 2017 19:36
  • Star 10 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
Star You must be signed in to star a gist
Save huyng/7969327 to your computer and use it in GitHub Desktop.
Place the following in your home directory at ~/.numpy-site.cfg in order to install numpy with openblas support
# Place the following in your home directory at ~/.numpy-site.cfg
# in order to install numpy with openblas support
[openblas]
libraries = openblas
library_dirs = /usr/local/opt/openblas/lib
include_dirs = /usr/local/opt/openblas/include
# Copyright (c) Huy Nguyen 2013
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without modification,
# are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright notice, this
# list of conditions and the following disclaimer.
#
# * Redistributions in binary form must reproduce the above copyright notice, this
# list of conditions and the following disclaimer in the documentation and/or
# other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
# ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
def _force_forder(x):
"""
Converts arrays x to fortran order. Returns
a tuple in the form (x, is_transposed).
"""
if x.flags.c_contiguous:
return (x.T, True)
else:
return (x, False)
def dot(A,B):
"""
Uses blas libraries directly to perform dot product
"""
from scipy import linalg
A, trans_a = _force_forder(A)
B, trans_b = _force_forder(B)
gemm_dot = linalg.get_blas_funcs("gemm", arrays=(A,B))
# gemm is implemented to compute: C = alpha*AB + beta*C
return gemm_dot(alpha=1.0, a=A, b=B, trans_a=trans_a, trans_b=trans_b)
@vbisin
Copy link

vbisin commented Mar 14, 2017

I have tried to run your fastdot code but am receiving an error (works fine with 2-tensors but not for 3-tensors). Once I edit the site.cfg file am I supposed to do something else (I suspect my openblas is not correctly synced with numpy)?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment