Openvino dev cloud for VPU and HDDLR

[Khurana__Tanisha]

I tried using Openvino for running a mask  rcnn script by using the foloowing plugin on dev cloud: 

qsub -l nodes=1:iei-mustang-v100-mx8 myscript.sh

from openvino.inference_engine import

IEPlugin plugin = IEPlugin(device="HDDLR")

And i got the following error; ValueError: Unknown plugin: HDDLR, expected one of: CPU,GPU,FPGA,MYRIAD,HETERO,HDDL,MULTI

Also I tried calculating fps for various devices on Intel dev cloud like GPU and CPU.

and when i used the command: !qsub -N testjob -l nodes=1:e3-1268l-v5 myjob

my average fps was around 0.4

while using this command : !qsub -N testjob -l nodes=e3-1268l-v5 myjob 

it came out to be around 3.5 fps.

This was on CPU i7 and xeon, whereas on GPU fps was around 0.7 .

What is the relevance of  nodes=1:e3-1268l-v5 as opposed to nodes=e3-1268l-v5 ?

[Munesh_Intel]

Hi Tanisha,

I am moving this topic to DevCloud for Edge forum for further assistance.

For future DevCloud related questions, I would suggest posting in the DevCloud forum.

Link : https://software.intel.com/en-us/forums/intel-devcloud-for-edge

Regards,

Munesh

[Eltablawy__Alaa]

Hi Tanisha,

Thanks for reaching out. For the first part of your question: The correct flag to target 8x VPU Myriad X is HDDL, not HDDLR, that's why you are seeing an error.

For the second part of your question: The correct way to submit a job to any compute node is "-l nodes=1:<target compute node tag>", removing the "1:" and using only the compute node tag is not a valid way.

Regards,

Alaa