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
1
vote
0 answers

shallow learning OpenBLAS warning -->cant model

I am trying to model using XGboost and an RF (separately and the code is terminating before the model is output.) I get the following error: OpenBLAS warning: precompiled NUM_THREADS exceeded, adding auxiliary array for thread metadata. Process…
Mike
  • 11
  • 1
1
vote
0 answers

A question about the efficiency of auto parallel openblas

I tried to use openblas to do a matrix multiplication and use its parallel version without writing DGEMM to different threads. The efficiency is not good when the common dimension is large. Is that a limitation of openblas that I should either write…
user235862
  • 11
  • 3
1
vote
0 answers

Does OpenBLAS use a fast matrix-multiplication algorithm?

I get 55 Gflops/s when multiplying two 10000 x 10000 matrices (counting 2 * 10000^3 flops for the entire computation). This was done on a single core, on a AMD Epyc 7313 running at 3.0GHz (boost clock turned off). Unless I am mistaken, the peak…
1
vote
0 answers

Stop numpy+sklearn from multithreading in google colab

My overall goal is to run a lot of instances of sklearn LogisticRegression on the same data in parallel using python multiprocessing.Pool. For this I compared the time of running unparalleled code with paralleled code. Here is the common prefix…
1
vote
1 answer

Python:: Sklearn OpenBLAS Error for KMeans

I get the following error when I try to use Kmeans: kmeans = KMeans(init='k-means++', n_clusters=numClusters, max_iter=100,…
P Ved
  • 109
  • 3
  • 13
1
vote
0 answers

Is there a way to tell R to use a specific .dll for OpenBLAS when I don't have read/write access to R directory?

I can follow the following instructions for my home computer fine. https://github.com/david-cortes/R-openblas-in-windows Which basically involves copying the compiled .dll for OpenBLAS into the R directory. But for work stuff, I am on a server where…
KevinM
  • 49
  • 5
1
vote
2 answers

CMAKE target link to a local openblas folder

I do not have Openblas installed. I do not want to install it. I want CMAKE to build my project using a local copy of OpenBLAS. I read this SO post: Using vendored OpenBLAS as a dependency in a CMake project Based on that, I made my CMAKELists.txt…
1
vote
0 answers

R tgp package installation under MKL environment

I am working on a server to which I don't have admin rights. I linked R to MKL for fast linear algebra using a Conda environment following these steps: Conda install r-essentials with MKL The linking of R to MKL was successful, the output for R…
BlS
  • 55
  • 4
1
vote
0 answers

Finding OpenBLAS on Windows with CMake

I followed the instructions here: https://github.com/xianyi/OpenBLAS/wiki/How-to-use-OpenBLAS-in-Microsoft-Visual-Studio for building and installing OpenBLAS, specifically 1. "Native (MSVC) ABI". And now all the includes and libraries and CMake…
Shadow0144
  • 123
  • 2
  • 9
1
vote
0 answers

How to fix this error for installing AutoGluon on Linux

I am trying to install AutoGluon on my university computer (Linux) I followed the instruction given here https://auto.gluon.ai/stable/index.html pip3 install -U pip pip3 install -U setuptools wheel # CPU version of pytorch has smaller footprint -…
asmgx
  • 7,328
  • 15
  • 82
  • 143
1
vote
0 answers

Spark netlib-java openBLAS

I am using word2vec in mllib and I want to improve the speed of linear operations in word2vec by configuring OpenBLAS. I configured OpenBLAS on the server by manually compiling OpenBLAS with the following command. git clone…
Rain Chen
  • 11
  • 1
1
vote
0 answers

numpy performance on multiple cpus

I am curious about numpy's performance on hpc machine with up to 128 cpus per node. I am running a dot product of two matrices, both have NxN elements, and progressively increasing the ncpus I request for this operation. Below is a graph showing the…
Sketos
  • 75
  • 8
1
vote
1 answer

Parallel Dense Matrix Diagonalization in C++

I have written a parallelized C++ code using OPENMP. I have been using EIGEN to deal with linear algebra. However, Eigen does not parallelize the eigen solver for dense matrix's. I' ve built OPENBLAS with openmp, however I do not see any performance…
1
vote
2 answers

openblas.bb:do_compile failed with exit code '1"

I am trying to build the openblas.bb in a yocto project but it fails. machine is "qemux86-64" DESCRIPTION = "OpenBLAS is an optimized BLAS library based on GotoBLAS2 1.13 BSD version." SUMMARY = "OpenBLAS : An optimized BLAS library" AUTHOR =…
A.A.
  • 125
  • 1
  • 11
1
vote
1 answer

compiling with MKL, OpenBLAS, and other libraries by using Makefile

I have tried to compile one probabilistic model posted on github (https://github.com/gerowam/influence), but for the person who is not familiar with C language and compiling, it is really hard to know how to install relevant libraries and do…
krcoder
  • 199
  • 2
  • 11