Parallel Processing using MYRIAD (Multiple Inference Requests)

[Catastrophe]

Let's say that I want to make use of 4 MYRIAD devices by making 4 inference requests in parallel. Is this possible? I was able to make 4 inference requests but I can't assign each inference request to each of the VPUs. The result, I am having 4x of the expected latency. 

Please help! Thanks!

[SuryaPSC_Intel]

Hi Catastrope,

Please refer to Multiple NCS devices for creating an OpenVINO™ toolkit application for multiple devices. Also you may refer to a similar thread for more information..

Best Regards,

Surya

[Catastrophe]

Thanks! But, is there an existing implementation of a sample source code in C++? 

[SuryaPSC_Intel]

Hi Catastrophe,

Please refer to Leverage Multiple Intel® Neural Compute Sticks with Intel® Distribution of OpenVINO™ Toolkit for more information,

Best Regards,

Surya