Regarding OpenAcc library in Jetson TX2

Hi All,

 We are trying to port our object recognition code on to NVIDIA TX2 platform and accelerate the performance using OpenACC library.  I need the link, where I can download OpenACC library on NVIDIA Jetson TX2 and the compiler information with which I can compile OpenACC code. Request help on this .

Hi,

PGI Community Edition doesn’t support ARM-based platform officially.
http://www.pgroup.com/products/community.htm?tum_campaign=CE&utm_source=dev_nvidia_oacc&utm_medium=web_link&utm_term=get_cd_text

But here is the sharing of OpenACC+TK1. You can give it a try.
http://scelementary.com/2015/04/30/openacc-on-jetson-tk1.html

Hi,

We are interested in using OpenACC on the TX2 and Xavier platforms.

Are there any updates on PGI support for these platforms in the past year? Is there a commercial/non-CE option?

Hi,

Just checking the PGI page. ARM support is still no available:
[url]https://www.pgroup.com/products/community.htm[/url]

Thanks.

Hi all,

have you already checked the OpenACC branch of GCC?

https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/heads/openacc-gcc-8-branch

One issue with gcc 8 is that it isn’t supported for nvcc. It’s too new. I think even gcc 7 is too new. There wouldn’t be anything stopping you from using it with non-CUDA apps.

Is there anything to look out for when running GPU accelerated apps using OpenACC while also running CUDA apps at the same time?

I’ve not used OpenACC before so I have no way to answer that (someone else here may know).

Thank you @linuxdev. I’m not sure, if we’re going to have the time to test it in our project, but I’ll post an update, if it comes to it.