Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.
Announcements
This community is designed for sharing of public information. Please do not share Intel or third-party confidential information here.
6592 Discussions

is there a MKL API for this use-case? R^TRx=b

Azua_Garcia__Giovann
177 Views

Hello,

I need to solve efficiently R^TRx=b where R is upper triangular. I cover this use-case doing two solve steps:

R^Ty=b,

Rx=y

and this works OK. However, I wonder whether there is a more efficient LAPACK or MKL API to solve the entire R^TRx=b at once.

Many TIA,

Best regards,

Giovanni

0 Kudos
4 Replies
Alexander_K_Intel2
177 Views
Hi, Just a few questions - are your matrix sparse or dense? And how you prepare this matrix - using MKL functionality or not? With best regards, Alexander Kalinkin
Azua_Garcia__Giovann
177 Views
Hi Alexander, The matrix is dense and upper triangular R. R is the result of up-dating and down-dating a Cholesky decomposition done as a series of incremental updates and not using MKL. However, the lower triangular is zero. Best regards, Giovanni
Alexander_K_Intel2
177 Views
Hi, You can solve your system by 2 call of dtrsv function. Is it suitable for you or you want to solve your equation by one call? With best regards, Alexander Kalinkin
Azua_Garcia__Giovann
177 Views
Hi Alexander, Exactly. My question is, is there a better way than making 2 calls of the dtrsv function. Or in other words, is it possible to solve R^TRx=b in one step. Thanks in advance, Best regards, Giovanni
Reply