AI accelerator support in oneAPI

[san__frank]

The ondAPI documentation mentions that oneAPI supports heterogenous architecture including CPU, GPU, FPGA and AI accelerator. Some public powerpoint have said AI accelerator is a future thing for oneAPI. Is there a roadmap for supporting AI accelerator from Intel, and from third party?

[ArunJ_Intel]

Hi Frank

We are forwarding your case for SME to check.

Thanks

Arun Jose

[JyotsnaK_Intel]

Hi Frank,

Thanks for showing you interest in AI Accelerator! Intel is working with AI accelerator teams both inside and outside Intel to incorporate AI accelerator support in the oneAPI specifications. We’ll have more to share about this later in 2020. 

[JyotsnaK_Intel]

Hi Frank,

Hope the answer helps. Do you have any other query which we I can help with? If not then I would like to close this thread. 

[san__frank]

Thanks Jyotsna. It would be good if you could let me know when there is any roadmap news regarding my question. I am keen to see how much interests the community have shown. If Intel could showcase with Habana roadmap committment in oneAPI, it would certainly help.

[JyotsnaK_Intel]

Hi Frank,

It is still under planning phase and we do not have any schedule to announce yet. I am closing this case. 

Thanks for showing your interest in it.