1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

OpenVINOを使ってWSLの速さを確認する

Last updated at Posted at 2019-06-26

目的

OpenVINOを使ってWSLの速さを確認する。

あまり良い速度検証の素材が思いついていないので、OpenVINOのデモを使いました。

検証環境

パソコン:ideapad 320S
CPU: Intel Core i3-7100U CPU @ 2.40GHz

OS: Windows10 バージョン1803(OSビルド 17134.829)
WSL(Windows Subsystem for Linux):上記のWindows10のとおり
Linux:Ubuntu 18.04LTS
表示:VcXsrv バージョン1.20.1.4
opencv:3.4.0
  ※3.4.1とかではないです。

利用したソフト:openvino_2019.1.087 ※1
デモソフト名:demo_security_barrier_camera
※1 WSL側は、あとから入れたので、openvino_2019.1.144

wsl_car.png

検証結果

inference(推論)の部分のみ。5回実行し、各々のfpsを示している。

--demo_security_barrier_camera
Windows  23.7  50.0  56.3  54.6  37.5fps
WSL  20.8  51.0  55.2  53.6  54.5fps

以下は、以前のWindows上の測定結果。
CPU FP32   22.6,  17.3fps

実行時出力抜粋
Windows

[ INFO ] InferenceEngine:
        API version ............ 1.6
        Build .................. 22443
[ INFO ] Parsing input parameters
[ INFO ] Capturing video streams from the video files or loading images
[ INFO ] Files were added: 1
[ INFO ]     C:\Program Files (x86)\IntelSWTools\openvino_2019.1.087\deployment_tools\demo\\car_1.bmp
[ INFO ] Number of input image files: 1
[ INFO ] Number of input video files: 0
[ INFO ] Number of input channels: 1
[ INFO ] Display resolution: 1920x1080
[ INFO ] Loading plugin CPU

        API version ............ 1.6
        Build .................. 22443
        Description ....... MKLDNNPlugin
[ INFO ] Loading network files for VehicleDetection
[ INFO ] Batch size is forced to  1
[ INFO ] Checking Vehicle Detection inputs
[ INFO ] Checking Vehicle Detection outputs
[ INFO ] Loading Vehicle Detection model to the CPU plugin
[ INFO ] Loading network files for VehicleAttribs
[ INFO ] Batch size is forced to 1 for Vehicle Attribs
[ INFO ] Checking VehicleAttribs inputs
[ INFO ] Checking Vehicle Attribs outputs
[ INFO ] Loading Vehicle Attribs model to the CPU plugin
[ INFO ] Loading network files for Licence Plate Recognition (LPR)
[ INFO ] Batch size is forced to  1 for LPR Network
[ INFO ] Checking LPR Network inputs
[ INFO ] Checking LPR Network outputs
[ INFO ] Loading LPR model to the CPU plugin
[ INFO ] Start inference




WSL

[ INFO ] InferenceEngine:
        API version ............ 1.6
        Build .................. custom_releases/2019/R1.1_28dfbfdd28954c4dfd2f94403dd8dfc1f411038b
[ INFO ] Parsing input parameters
[ INFO ] Capturing video streams from the video files or loading images
[ INFO ] Files were added: 1
[ INFO ]     /opt/intel/openvino_2019.1.144/deployment_tools/demo/car_1.bmp
[ INFO ] Number of input image files: 1
[ INFO ] Number of input video files: 0
[ INFO ] Number of input channels: 1
[ INFO ] Display resolution: 1920x1080
[ INFO ] Loading plugin CPU

        API version ............ 1.6
        Build .................. 23780
        Description ....... MKLDNNPlugin
[ INFO ] Loading network files for VehicleDetection
[ INFO ] Batch size is forced to  1
[ INFO ] Checking Vehicle Detection inputs
[ INFO ] Checking Vehicle Detection outputs
[ INFO ] Loading Vehicle Detection model to the CPU plugin
[ INFO ] Loading network files for VehicleAttribs
[ INFO ] Batch size is forced to 1 for Vehicle Attribs
[ INFO ] Checking VehicleAttribs inputs
[ INFO ] Checking Vehicle Attribs outputs
[ INFO ] Loading Vehicle Attribs model to the CPU plugin
[ INFO ] Loading network files for Licence Plate Recognition (LPR)
[ INFO ] Batch size is forced to  1 for LPR Network
[ INFO ] Checking LPR Network inputs
[ INFO ] Checking LPR Network outputs
[ INFO ] Loading LPR model to the CPU plugin
[ INFO ] Start inference

検証結果(追加)

もう一度、計ってみました。

inference(推論)の部分のみ。各々のfpsを示している。
[]内の数字は、測定を実施した順番。
(13回目までは、ノートPCの電源は未接続。)

--demo_security_barrier_camera
Windows  
20.66[1], 16.72[5], 26.71[6], 55.72[7], 48.82[8], 54.99[9],
21.86[17], 53.89[18], 54.09[19]fps

WSL  
2.88[2], 12.06[3], 19.34[4], 18.39[10], 17.68[11], 22.25[12], 18.32[13],
21.54[14], 20.52[15], 18.94[16]fps

⇒ 今日は、WSLは、全く、やる気がないようです。
  (いい加減な検証で、すみません。)

補足(動作させるための作業メモ)

以下のサイトに記載のコマンドのとおり実施した。
https://docs.openvinotoolkit.org/latest/_docs_install_guides_installing_openvino_linux.html#install-openvino

ビルドで以下のようなエラーが(大量に)出た

/opt/intel/openvino_2019.1.144/opencv/lib/libopencv_highgui.so.4.1.0: undefined reference to `gtk_main_iteration'

以下の対応を実施した(いろいろな対処を試したので、下記が正しい対処かは不明)

$ sudo apt-get install libgtk-3-dev

まとめ

Windowsでの実行と、WSL上での実行で、速度差はない感じ。
そもそも、測定毎の大きなバラツキのほうが問題。
もう少し、まともな測定方法を考える必要あり。
以前の結果に比して、極端に速くなっている理由不明。
ということで、
問題しかない、まとめです。

今後

コメントなどあればお願いします。:candy:

関連

OpenVINOを使ってCPU内蔵GPUの速さを確認する

1
1
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?