Forums - could NPE auto select the runtime? Or should manually tell NPE run in only one runtime?

2 posts / 0 new
Last post
could NPE auto select the runtime? Or should manually tell NPE run in only one runtime?
luntingh
Join Date: 23 Jun 21
Posts: 1
Posted: Thu, 2021-06-24 02:24

NPE can use CPU, GPU, DSP and so on.

So, I am confued about the runtime usage when doing infer, such as use VGG16 to do classification.

Could NPE auto select the runtime?

 i.e., when doing VGG16 infer, NPE can use all the computing resource at the same time?

Or NPE can only use one computing resource at the same time? such as only use GPU.

Or NPE can auto decide which computing should be used?

Or user should make decision?

If i want to use all the resource at the same time to do VGG16 inference, how?

 

thanks and best regards

 

 

  • Up0
  • Down0
ap.arunraj
Join Date: 20 Apr 20
Posts: 21
Posted: Tue, 2021-07-13 05:29

Hello luntingh,
If you're using snpe-net-run, by default it uses CPU as the runtime. Also you can specify the runtime order by using 
--runtime_order <VAL,VAL,VAL>, where VAL can be any of the supported runtime specified here
Incase you're building a C++ application that uses SNPE, you can use 
setRuntimeProcessorOrder method of SNPE builder class to define the order of precedence for runtime.

  • 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.