[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Help-gsl] SVD with GSL outperformed by MatLab by a factor of 10

From: Kalb Max
Subject: [Help-gsl] SVD with GSL outperformed by MatLab by a factor of 10
Date: Tue, 14 Dec 2010 13:49:56 +0100
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv: Gecko/20100915 Thunderbird/3.1.4


we compared the GSL-svd to MatLab and found an order of magnitude in time performance in favor of MAtLab nearly independent of the matrix sizes. We tried 100x52 up to 2000x52 as dimensions of the real matrices, which are not sparse. For a 800x 52 matrix we have 210 ms with GSL and 7 ms with MatLab for the svd for the following commands:

gsl_linalg_SV_decomp_mod(Atempm, X, Vm, S, work)
gsl_blas_dgemm(CblasNoTrans, CblasNoTrans, 1.0, Vm, SiD, 0.0, VSiD);
gsl_blas_dgemm(CblasNoTrans, CblasTrans, 1.0, VSiD, Atempm, 0.0, MoIm);

We work with:
Intel(R) Core(TM)2 Duo CPU E6850 @ 3.00 GHz, 2 GB RAM
WinXP S3
Visual Studio 2005
GSL 1.12

We could live with less precise singular values, since we throw away the smallest ones. Additionally, we do not need the according vectors of the U-matrix.

Any idead how to improve the GSL version?

Thanks a lot!


Attachment: max_kalb.vcf
Description: Vcard

reply via email to

[Prev in Thread] Current Thread [Next in Thread]