[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Using matrix decompositions
From: |
Patrick Alken |
Subject: |
Re: Using matrix decompositions |
Date: |
Sat, 30 Jan 2021 09:59:20 -0700 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.10.0 |
Hello,
There are 2 interfaces for the QR decomposition, the original
interface which uses Level-2 BLAS, and a recently added interface using
Level-3 BLAS (_r suffix to function names). In both cases, there is an
"unpack" function to extract the Q matrix:
gsl_linalg_QR_unpack
gsl_linalg_QR_unpack_r
The R matrix is always stored in the upper triangle of the input matrix
(A) by the QR_decomp routines.
In most applications, you never need to actually construct the full Q
matrix itself. If you describe more about what you want to do, or post
your source code, we can probably help more.
Thanks,
Patrick
On 1/30/21 1:49 AM, Iaroslav Postovalov wrote:
> Hello,
>
> I'm quite confused with using QR decomposition with GSL. The library
> returns data in very unclear data format, and I don't understand how to get
> Q and R matrices from it.
>
> Certain users may have the same problem for other decompositions (since
> literally all of them require some effort to extract the resulting
> matrices), and I think that the GSL documentation has to provide some
> instructions for them.
>
> All the bests,
> Iaroslav