Questions tagged [openblas]

Questions regarding OpenBLAS, an open source library that implements BLAS API and has hand-crafted optimizations for specific processor types.

OpenBLAS is an open source implementation of BLAS, the Basic Linear Algebra Subprograms library, optimized for high performance on multi-core x86-64 processors. It was forked from the GotoBLAS2-1.13 BSD version and was tuned to have performance similar to Intel MKL.

277 questions
7
votes
2 answers

Why is matrix product slower when matrix has very small values?

I create two matrices A and B of the same dimension. A contains larger values than B. The matrix multiplication A %*% A is about 10 times faster than B %*% B. Why is this? ## disable openMP library(RhpcBLASctl); blas_set_num_threads(1);…
Nairolf
  • 2,418
  • 20
  • 34
7
votes
4 answers

Tutorial for installing numpy with OpenBLAS on Windows

Please, I do need a light here. I want to install numpy using a good BLAS/LAPACK lib on Windows, but absolutely no page explains the process well enough. It seems OpenBLAS is a good and fast option. The goal is to use "theano" with "keras", and…
Daniel Möller
  • 84,878
  • 18
  • 192
  • 214
6
votes
2 answers

macOS Monterey - Install Scipy Error "No lapack/blas resources found"

I have installed pyenv on Mac using brew install pyenv. I want to run a python project which requires installation of scipy==1.5.4. I tried installing it using the command pip3 install scipy==1.5.4 but I got an error No lapack/blas resources…
Lopez
  • 461
  • 5
  • 19
6
votes
0 answers

Why can R be linked to a shared BLAS later even if it was built with `--with-blas = lblas`?

The BLAS section in R installation and administration manual says that when R is built from source, with configuration parameter --without-blas, it will build Netlib's reference BLAS into a standalone shared library at R_HOME/lib/libRblas.so, along…
Zheyuan Li
  • 71,365
  • 17
  • 180
  • 248
6
votes
1 answer

installing theano "blas error"

>>> import numpy # OK! >>> import scipy # OK! >>> import theano # warning WARNING (theano.tensor.blas): Failed to import scipy.linalg.blas, and Theano flag blas.ldflags is empty. Falling back on slower implementations for dot(matrix, vector),…
babeyh
  • 659
  • 2
  • 7
  • 19
6
votes
3 answers

"/usr/bin/ld: cannot find -lopenblas" error in Caffe compilation

When I was compiling Caffe, I had this error, despite OpenBLAS is installed: AR -o .build_release/lib/libcaffe.a LD -o .build_release/lib/libcaffe.so /usr/bin/ld: cannot find -lopenblas collect2: ld devolvió el estado de salida 1 make: ***…
6
votes
1 answer

Why is numpy.dot much faster than numpy.einsum?

I have numpy compiled with OpenBlas and I am wondering why einsum is much slower than dot (I understand in the 3 indices case, but I dont understand why it is also less performant in the two indices case)? Here an example: import numpy as np A =…
varantir
  • 6,624
  • 6
  • 36
  • 57
6
votes
1 answer

How to force armadillo library to link to local OpenBLAS during compilation

I need to have armadillo (current version is 5.100.1) available as a local library within $HOME (cluster application, and can't install on every compute node, but $HOME is shared folder). I'm using cmake to manage the application, and have been able…
Rolf
  • 139
  • 1
  • 9
5
votes
0 answers

How can I have a finer control on number of threads used for each BLAS kernel call on CPU?

I am writing an OpenMP code calling different BLAS kernels, mostly DGEMMs with different sizes, in different threads. To maximize performance I want to have control over the number of threads I am calling for each BLAS. It seems that it is a very…
Aznaveh
  • 558
  • 8
  • 27
5
votes
0 answers

Checking BLAS and LAPACK libraries links in scipy and numpy

I'm trying to link my code (scipy.linalg.solve notably) to a given library like OpenBLAS. I am using python 3.7.7 in a virtual environment: Python 3.7.7 (default, Mar 29 2020, 18:12:06) [Clang 11.0.3 (clang-1103.0.32.29)] on darwin Also of note, I…
William Abma
  • 415
  • 3
  • 14
5
votes
3 answers

Setting up Anaconda for AMD Ryzen without MKL

Like many others, I've bought myself a new Ryzen CPU. I need to use Anaconda Python for my PhD (together with Tensorflow etc). Since Anaconda now comes pre-packaged with MKL which is slow on AMD CPUs, what is the best way to setup an Anaconda…
stevew
  • 674
  • 10
  • 19
5
votes
1 answer

CMake, .dll vs .dll.a vs .a: What library should I specify in the CMake configuration?

I am working on Windows 10 and CMake 3.14.1. I am trying to build OpenCV (Version 4.0.1) myself, since I need the contrib modules and the SFM module. I downloaded a prebuilt version of openBLAS 0.2.19. It has 3 directories, bin, include, and…
jiggy
  • 262
  • 2
  • 12
5
votes
1 answer

OpenBLAS issue with ComBat() function of the R Bioconductor SVA package on Torque cluster

I've a problem with ComBat() function from the SVA package of Bioconductor in R. On my laptop (Latitude 5590 running a Linux Ubuntu 18 operating system), it works well. But if I ran it on a TORQUE cluster, the call to the ComBat() function generates…
DavideChicco.it
  • 3,318
  • 13
  • 56
  • 84
5
votes
1 answer

Numpy dot operation is not using all cpu cores

I am doing numpy dot product on two matrices (Let us assume a and b are two matrices). When the shape of a is (10000, 10000) and shape of b is (1, 10000) then the numpy.dot(a, b.T) is using all the CPU cores. But when the shape of a is (10000,…
Ram Idavalapati
  • 666
  • 1
  • 10
  • 22
5
votes
2 answers

Installing OpenBLAS on CentOS / Fedora

In a Java project I am using matrix-toolkits-java (MTJ) for efficient matrix multiplication. This relies on netlib-java, which in turn relies on an optimized implementation of BLAS and LAPACK installed on the machine. It specifically looks for…
TTT
  • 6,505
  • 10
  • 56
  • 82
1
2
3
18 19