Questions tagged [blas]
The Basic Linear Algebra Subprograms are a standard set of interfaces for low-level vector and matrix operations commonly used in scientific computing.
blas
918
questions
194
votes
4
answers
52k
views
What is the relation between BLAS, LAPACK and ATLAS
I don't understand how BLAS, LAPACK and ATLAS are related and how I should use them together! I have been looking through all of their manuals and I have a general idea of BLAS and LAPACK and how to ...
- 3,946
155
votes
8
answers
61k
views
How does BLAS get such extreme performance?
Out of curiosity I decided to benchmark my own matrix multiplication function versus the BLAS implementation... I was to say the least surprised at the result:
Custom Implementation, 10 trials of
...
- 6,071
141
votes
5
answers
95k
views
How to check BLAS/LAPACK linkage in NumPy and SciPy?
I am builing my numpy/scipy environment based on blas and lapack more or less based on this walk through.
When I am done, how can I check, that my numpy/scipy functions really do use the previously ...
- 13.8k
140
votes
3
answers
49k
views
Why does multiprocessing use only a single core after I import numpy?
I am not sure whether this counts more as an OS issue, but I thought I would ask here in case anyone has some insight from the Python end of things.
I've been trying to parallelise a CPU-heavy for ...
- 73.1k
119
votes
5
answers
47k
views
Benchmarking (python vs. c++ using BLAS) and (numpy)
I would like to write a program that makes extensive use of BLAS and LAPACK linear algebra functionalities. Since performance is an issue I did some benchmarking and would like know, if the approach I ...
- 13.8k
86
votes
2
answers
3k
views
Is armadillo solve() thread safe?
In my code I have loop in which I construct and over determined linear system and try to solve it:
#pragma omp parallel for
for (int i = 0; i < n[0]+1; i++) {
for (int j = 0; j < n[1]+1; j++...
- 1,066
83
votes
10
answers
67k
views
MatLab error: cannot open with static TLS
Since a couple of days, I constantly receive the same error while using MATLAB which happens at some point with dlopen. I am pretty new to MATLAB, and that is why I don't know what to do. Google doesn'...
- 931
76
votes
16
answers
64k
views
TensorFlow: InternalError: Blas SGEMM launch failed
When I run sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys}) I get InternalError: Blas SGEMM launch failed. Here is the full error and stack trace:
InternalErrorTraceback (most recent call ...
- 2,579
68
votes
1
answer
3k
views
Distributing Cython based extensions using LAPACK
I am writing a Python module that includes Cython extensions and uses LAPACK (and BLAS). I am open to using either clapack or lapacke, or some kind of f2c or f2py solution if necessary. What is ...
- 4,011
66
votes
20
answers
78k
views
TensorFlow: Blas GEMM launch failed
When I'm trying to use TensorFlow with Keras using the gpu, I'm getting this error message:
C:\Users\nicol\Anaconda3\envs\tensorflow\lib\site-packages\ipykernel\__main__.py:2: UserWarning: Update ...
- 709
57
votes
3
answers
14k
views
Purpose of LDA argument in BLAS dgemm?
The Fortran reference implementation documentation states:
* LDA - INTEGER.
* On entry, LDA specifies the first dimension of A as declared
* in the calling (sub) program. When ...
- 27.8k
54
votes
3
answers
49k
views
Compiling numpy with OpenBLAS integration
I am trying to install numpy with OpenBLAS , however I am at loss as to how the site.cfg file needs to be written.
When the installation procedure was followed the installation completed without ...
- 849
50
votes
4
answers
36k
views
multithreaded blas in python/numpy
I am trying to implement a large number of matrix-matrix multiplications in Python. Initially, I assumed that NumPy would use automatically my threaded BLAS libraries since I built it against those ...
- 938
46
votes
1
answer
7k
views
Keras not using multiple cores
Based on the famous check_blas.py script, I wrote this one to check that theano can in fact use multiple cores:
import os
os.environ['MKL_NUM_THREADS'] = '8'
os.environ['GOTO_NUM_THREADS'] = '8'
os....
- 5,431
40
votes
3
answers
36k
views
Find out if/which BLAS library is used by Numpy
I use numpy and scipy in different environments (MacOS, Ubuntu, RedHat).
Usually I install numpy by using the package manager that is available (e.g., mac ports, apt, yum).
However, if you don't ...
- 579
39
votes
6
answers
13k
views
Linking Intel's Math Kernel Library (MKL) to R on Windows
Using an alternative BLAS for R has several advantages, see e.g. https://cran.r-project.org/web/packages/gcbd/vignettes/gcbd.pdf.
Microsoft R Open https://mran.revolutionanalytics.com/documents/rro/...
- 7,911
28
votes
3
answers
14k
views
BLAS: gemm vs. gemv
Why does BLAS have a gemm function for matrix-matrix multiplication and a separate gemv function for matrix-vector multiplication? Isn't matrix-vector multiplication just a special case of matrix-...
- 68k
28
votes
2
answers
9k
views
Multi-threaded integer matrix multiplication in NumPy/SciPy
Doing something like
import numpy as np
a = np.random.rand(10**4, 10**4)
b = np.dot(a, a)
uses multiple cores, and it runs nicely.
The elements in a, though, are 64-bit floats (or 32-bit in 32-bit ...
- 2,226
27
votes
5
answers
6k
views
Fastest way to negate a std::vector
Assume I have a std::vector of double, namely
std::vector<double> MyVec(N);
Where N is so big that performance matters. Now assume that MyVec is a nontrivial vector (i.e. it is not a vector of ...
- 949
27
votes
6
answers
7k
views
Running Scipy on Heroku
I got Numpy and Matplotlib running on Heroku, and I'm trying to install Scipy as well. However, Scipy requires BLAS[1] to install, which is not presented on the Heroku platform. After contacting ...
- 321
27
votes
3
answers
8k
views
calling dot products and linear algebra operations in Cython?
I'm trying to use dot products, matrix inversion and other basic linear algebra operations that are available in numpy from Cython. Functions like numpy.linalg.inv (inversion), numpy.dot (dot product),...
user248237
26
votes
4
answers
10k
views
R detection of Blas version
Is there a way of detecting the version of BLAS that R is using from inside R? I am using Ubuntu, and I have a couple of BLAS versions installed - I just don't know which one is "active" from R's ...
- 3,885
24
votes
2
answers
21k
views
Link ATLAS/MKL to an installed Numpy
TL;DR how to link ATLAS/MKL to existing Numpy without rebuilding.
I have used Numpy to calculate with the large matrix and I found that it is very slow because Numpy only use 1 core to do calculation....
- 653
24
votes
5
answers
27k
views
How to make sure the numpy BLAS libraries are available as dynamically-loadable libraries?
The theano installation documentation states, that theano will as a default use the BLAS libraries from numpy, if the "BLAS libraries are available as dynamically-loadable libraries". This seems not ...
- 34.3k
24
votes
1
answer
2k
views
Implementing faster python inner product with BLAS
I found this useful tutorial on using low-level BLAS functions (implemented in Cython) to get big speed improvements over standard numpy linear algebra routines in python. Now, I've successfully ...
- 2,954
22
votes
1
answer
2k
views
Replicating BLAS matrix multiplication performance: Can I match it?
Background
If you have been following my posts, I am attempting to replicate the results found in Kazushige Goto's seminal paper on square matrix multiplication C = AB. My last post regarding this ...
- 589
21
votes
1
answer
38k
views
Installing LAPACK and BLAS Libraries for C on Mac OS
I wanted instructions/websites from where I could download LAPACK and BLAS libraries for use in my C programs. I also wanted to know how I could link these to the gcc compiler from terminal.
- 473
19
votes
4
answers
15k
views
Element-wise vector-vector multiplication in BLAS?
Is there a means to do element-wise vector-vector multiplication with BLAS, GSL or any other high performance library ?
- 1,070
18
votes
2
answers
13k
views
Numpy, BLAS and CUBLAS
Numpy can be "linked/compiled" against different BLAS implementations (MKL, ACML, ATLAS, GotoBlas, etc). That's not always straightforward to configure but it is possible.
Is it also possible to "...
- 17.4k
18
votes
2
answers
5k
views
Mystified by qr.Q(): what is an orthonormal matrix in "compact" form?
R has a qr() function, which performs QR decomposition using either LINPACK or LAPACK (in my experience, the latter is 5% faster). The main object returned is a matrix "qr" that contains in the upper ...
- 10.2k
18
votes
4
answers
20k
views
What is the BigO of linear regression?
How large a system is it reasonable to attempt to do a linear regression on?
Specifically: I have a system with ~300K sample points and ~1200 linear terms. Is this computationally feasible?
- 76.9k
18
votes
1
answer
30k
views
how to check if BLAS and ATLAS already installed
I'm trying to install armadillo library onto my linux system(ubuntu 12.04). The BOOST BLAS ATLAS and LAPACK is required first for the installation. Is there a way to check if those libraries are ...
- 2,187
17
votes
2
answers
12k
views
Set max number of threads at runtime on numpy/openblas
I'd like to know if it's possible to change at (Python) runtime the maximum number of threads used by OpenBLAS behind numpy?
I know it's possible to set it before running the interpreter through the ...
- 3,320
17
votes
3
answers
6k
views
Why does MATLAB/Octave wipe the floor with C++ in Eigenvalue Problems?
I'm hoping that the answer to the question in the title is that I'm doing something stupid!
Here is the problem. I want to compute all the eigenvalues and eigenvectors of a real, symmetric matrix. I ...
- 1,668
16
votes
4
answers
16k
views
What is a good free (open source) BLAS/LAPACK library for .net (C#)? [closed]
I have a project written in C# where I need to do various linear algebraic operations on matrices (like LU-factorization).
Since the program is mainly a prototype created to confirm a theory, a C# ...
- 15.3k
16
votes
2
answers
2k
views
No speedup for vector sums with threading
I have a C++ program which basically performs some matrix calculations. For these I use LAPACK/BLAS and usually link to the MKL or ACML depending on the platform. A lot of these matrix calculations ...
- 173
16
votes
1
answer
14k
views
Fortran 90/95 library for sparse matrices?
I am looking for a library for dealing with sparse matrices in fortran 90/95. I only need very basic operations like matrix-vector multiplication. What do you suggest I use?
I have searched around ...
- 707
16
votes
2
answers
7k
views
Any good documentation for the cblas interface? [closed]
Can someone recommend a good reference or tutorial for the cblas interface? Nothing comes up on google, all of the man pages I've found are for the fortran blas interface, and the pdf that came with ...
- 23.7k
15
votes
3
answers
13k
views
How to use numpy with OpenBLAS instead of Atlas in Ubuntu?
I have looked for an easy way to install/compile Numpy with OpenBLAS but didn't find an easy answer. All the documentation I have seen takes too much knowledge as granted for someone like me who is ...
- 675
15
votes
3
answers
10k
views
Theano CNN on CPU: AbstractConv2d Theano optimization failed
I'm trying to train a CNN for object detection on images with the CIFAR10 dataset for a seminar at my university but I get the following Error:
AssertionError: AbstractConv2d Theano optimization ...
- 293
15
votes
1
answer
2k
views
cholesky decomposition ScaLapack error
I'm getting the following error and i'm not sure why.
{ 1, 1}: On entry to PDPOTRF parameter number 2 had an illegal value
{ 1, 0}: On entry to PDPOTRF parameter number 2 had an ...
- 12k
14
votes
2
answers
20k
views
"Attempting to perform BLAS operation using StreamExecutor without BLAS support" error occurs
my computer has only 1 GPU.
Below is what I get the result by entering someone's code
[name: "/device:CPU:0" device_type: "CPU" memory_limit: 268435456
locality {} incarnation: ...
- 231
14
votes
1
answer
5k
views
Why are there no BLAS routines for addition and subtraction
In BLAS there are routines like
dscal scale a vector by a constant
dinit initialize a vector with given value
daxpy perform y = a*x + y
and so on. But there are apparently no routines ...
- 5,743
14
votes
1
answer
535
views
Faster evaluation of matrix multiplication from right to left
I noticed that evaluating matrix operations in quadratic form from right to left is significantly faster than left to right in R, depending on how the parentheses are placed. Obviously they both ...
- 273
13
votes
4
answers
23k
views
Numpy Pure Functions for performance, caching
I'm writing some moderately performance critical code in numpy.
This code will be in the inner most loop, of a computation that's run time is measured in hours.
A quick calculation suggest that this ...
- 27.7k
13
votes
3
answers
4k
views
Is it possible to switch between BLAS libraries without recompiling program?
For example can I have Atlas, OpenBlas, MKL installed on my Ubuntu 14.04 at the same time and switch between them without recompiling Caffe?
- 20.8k
12
votes
5
answers
18k
views
libgfortran: version `GFORTRAN_1.4' not found
I am getting the following error when I trying to a run mex file in MATLAB:
??? Invalid MEX-file
'findimps3.mexa64':
/MATLAB/bin/glnxa64/../../sys/os/glnxa64/libgfortran.so.3: version `GFORTRAN_1.4' ...
- 4,656
12
votes
3
answers
1k
views
How can I determine which matrix libraries my R install is using?
I am having a matrix error when using the computer cluster at my university that I cannot reproduce on my local machine. I think it might be due to a difference of matrix libraries (BLAS, LAPACK, ...
- 1,160
12
votes
1
answer
3k
views
Floating point math in python / numpy not reproducible across machines
Comparing the results of a floating point computation across a couple of different machines, they are consistently producing different results. Here is a stripped down example that reproduces the ...
- 705
12
votes
1
answer
3k
views
performance of NumPy with different BLAS implementations
I'm running an algorithm that is implemented in Python and uses NumPy. The most computationally expensive part of the algorithm involves solving a set of linear systems (i.e. a call to numpy.linalg....
- 1,543