Forums - Timing improvement of qmath lib qm_vqf32_sqrt

3 posts / 0 new
Last post
Timing improvement of qmath lib qm_vqf32_sqrt
ANN
Join Date: 21 Dec 16
Posts: 22
Posted: Thu, 2018-03-15 23:23

Hello,

I am refeering to Hexagon cDSP on SDM660.

When calculating square root on cDSP, can we obtain a time gain using qm_vqf32_sqrt()function form qmathlib over sqrt() function from standard C Lib?

Thank you.

  • Up0
  • Down0
e.raju789
Join Date: 9 Mar 18
Posts: 5
Posted: Fri, 2018-03-23 05:51

if you have multiple elements to process in parallel (e.g. 32 or more) and have other similar math operations to perform, that function will bring significant gain.

There is a cost to converting from native float to the pseudo float format used in qmath and back.

Regards, 

Chenna 

  • Up0
  • Down0
tom.zheng
Join Date: 11 Jun 18
Posts: 3
Posted: Wed, 2018-06-13 03:03

Hi Chenna,

Have you got an example of ultilizing the qmath library while having vectors with more than 32 in size for math operations?

BR,

Tom

  • Up0
  • Down0
or Register

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.