Intel® Distribution of OpenVINO™ Toolkit
Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.

Performance comparison: cpu vs ncs2

ShihSheng__Chen
Beginner
523 Views

Hi

I've tested demo_benchmark_app of the OpenVINO examples provided by Intel on both the Myriad X(NCS2) and Intel(R) Core(TM) i5-6500 CPU.The result is below:

//////////////////////////////////////////////////////////////////////////////////////////////////////

Full device name: Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz

Count:      1000 iterations
Duration:   2569.53 ms
Latency:    10.01 ms
Throughput: 389.18 FPS
//////////////////////////////////////////////////////////////////////////////////////////////////////

Full device name: Intel Movidius Myriad X VPU

Count:      1000 iterations
Duration:   3782.36 ms
Latency:    15.07 ms
Throughput: 264.39 FPS

//////////////////////////////////////////////////////////////////////////////////////////////////////

 

So the i5-6500 CPU is faster than Myriad X(NCS2)? Is this result normal?

 

Best Regards,

Jerry Chen

 

0 Kudos
3 Replies
nikos1
Valued Contributor I
523 Views

Looks normal to me. I get similar results.

What was your expectation?

Cheers,

nikos

 

MAURICIOAL_R_Intel
Moderator
523 Views

Thank you for reaching out, I apologize for the delay in my response.

Yes, you are right. The benchmarks showed in this performance documentation, demonstrate high-performance gains on several public neural networks for a streamlined, quick deployment on Intel® CPU, VPU and FPGA platforms. You can use this information to help you decide which hardware is best for your applications and solutions.

Regards,

Mauricio R.

TA1
Beginner
361 Views

Hello I'm beginner , sorry for poor English

I got similar (NCS2 was much later than CPU)results, but not demo models.

I made a deep learnning models which predict MNIST, and then converted frozen .pb file to IR format(.bin .xml) and inferred using OpenVINO.

 

Is this result from USB 2.0 of my PC?

Is it depends on complexity of model(MNIST is so simple?)?

 SPEC bellw:

python 3.6.8

OpenVINO 2021.2

Tensorflow 2.4.0

Windows 10

Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz 

Reply