How to use the OpenCL compiler of OpenVINO?

[Chen__Oliver]

Hi!

I'm considering the possibility of running OpenCL on the SHAVE processor of Intel Neural Compute Stick 2 (NCS2) connected to an ARM board (Dragonboard 410C). The target is use OpenCL to accelerate OpenCV applications, or even to other general purpose ones.

In this page: Release Notes for Intel® Distribution of OpenVINO™ toolkit 2019

There is a description about an OpenCL compiler:

HDDL Plugin

Added OpenCL™ custom layer support for the Intel® Vision Accelerator Design with Intel® Movidius™ VPUs in the Preview Feature mode:

OpenCL™ compiler, targeting SHAVE processor only, is redistributed with OpenVINO™. OpenCL support is provided by ComputeAorta, and is distributed under a license agreement between Intel and Codeplay Software Ltd.

I'm wondering if there is any other information about how to use this OpenCL compiler?

Regards,

Oliver

[David_C_Intel]

Hi Oliver, 

Thank you for reaching out. This guide might be useful for your needs.

If you need additional help, feel free to post here again.

Best regards, 

David

[Chen__Oliver]

Hi David,

Thank you for your reply and providing the link to that document.

Thanks to that guide, I've found the directory of OpenCL tools in the installed OpenVINO:

  <INSTALL_DIR>/deployment_tools/tools/cl_compiler

In that directory there are few binary executable files, which I believe are compiled for Intel x86 CPU.

As I want to use OpenVINO on ARM (the Dragonboard 410c board), I'm wondering whether I can:

1. find the source code of those OpenCL tools to compile for ARM, just like I compiled the Deep Learning Deployment Toolkit (DLDT) for ARM? or

2. find the binary executable of those OpenCL tools compiled for ARM?

[David_C_Intel · Accepted solution]

Hello Oliver

Thank you for replying.

Looking through the DLDT repository under tools, we can see that OpenCL is not available. In fact, the only way to use OpenVINO™ toolkit on ARM is with DLDT, and it does not include OpenCL. 

Regards, 

David

[Chen__Oliver]

Hi David,

OK, now I understand.

Thanks for your reply and explanation!