Forums - Float 16 in IDL

3 posts / 0 new
Last post
Float 16 in IDL
vijay.alagappan
Join Date: 15 Feb 24
Posts: 3
Posted: Wed, 2024-05-22 22:39

I'm trying to pass float16 array from the android HLOS side to the DSP side. I have used the IDL before for defining the interfaces and was able to do it for float32, int16, int8, etc. But I'm not sure how to use float 16 on the IDL. I tried using __fp16, but it throws up an error.

The table on the IDL documentation in the Hexagon SDK 5.5.0.1 shows IDL equivalents for C types, but I couldn't find any information on the float16 type. Is it possible to use float 16 in the IDL? I have float16 arrays on the android CPU side and want to pass it to the DSP side to make use of DSP compute for float16s. Kindly provide me some information on this. Thanks.

 

  • Up0
  • Down0
prabpand
Join Date: 14 May 19
Posts: 3
Posted: Fri, 2024-05-24 01:08

Hi Vijay, 

IDL doesn't have float16 support. It only supports "float" datatype. Do you see any issues while using that? 


Regards,

Prabha

  • Up0
  • Down0
vijay.alagappan
Join Date: 15 Feb 24
Posts: 3
Posted: Fri, 2024-05-24 03:43

Thanks a lot for the information. (I was looking to use float16 because my inputs were in float16 and converting to float32 would have added some overhead., I'll now try some workarounds)

 

  • 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.