oneVPL evaluation on devcloud

Jitendra_S_Intel · ‎08-19-2021

I'm trying to use oneAPI DevCloud to test some oneVPL samples. I have downloaded sample video from https://github.com/intel-iot-devkit/sample-videos/blob/master/people-detection.mp4.

I also built the oneVPL samples from login node and tried to run them on a compute node by using an interactive shell. Using following command to log into the compute node:

qsub -I -l nodes=1:quad_gpu:ppn=2 -d .

The decode and VPP samples don't seem to run properly. Please check.

SW based decode attempt below. The video is 49 seconds long @ 12fps. The output shows 9 frames "processed".

uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$ time ./hello-vpp -sw -i ./test.h264 -w 768 -h 432
Implementation details:
  ApiVersion:           2.4
  Implementation type:  SW
  AccelerationMode via: NA
  Path: /glob/development-tools/versions/oneapi/2021.3/inteloneapi/vpl/2021.4.0/lib/libvplswref64.so.1

Processing ./test.h264 -> out.raw
Processed 9 frames

real    0m4.939s
user    0m0.019s
sys     0m0.052s
uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$

HW based decode attempt:

uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$ time ./hello-vpp -hw -i ./test.h264 -w 768 -h 432
Cannot create session -- no implementations meet selection criteria
Processed 0 frames

real    0m0.027s
user    0m0.009s
sys     0m0.024s
uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$ cd ..

uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-decode/build$ ./hello-decode -sw -i ../../hello-vpp/build/test.h264
Implementation details:
  ApiVersion:           2.4
  Implementation type:  SW
  AccelerationMode via: NA
  Path: /glob/development-tools/versions/oneapi/2021.3/inteloneapi/vpl/2021.4.0/lib/libvplswref64.so.1

Error initializing decode

Decoded 0 frames
uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-decode/build$

Input file information:

uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$ ffmpeg -i ./test.h264 -f null -
ffmpeg version 4.2.4-1ubuntu0.1 Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 9 (Ubuntu 9.3.0-10ubuntu2)
  configuration: --prefix=/usr --extra-version=1ubuntu0.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-avisynth --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-nvenc --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 31.100 / 56. 31.100
  libavcodec     58. 54.100 / 58. 54.100
  libavformat    58. 29.100 / 58. 29.100
  libavdevice    58.  8.100 / 58.  8.100
  libavfilter     7. 57.100 /  7. 57.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  5.100 /  5.  5.100
  libswresample   3.  5.100 /  3.  5.100
  libpostproc    55.  5.100 / 55.  5.100
Input #0, h264, from './test.h264':
  Duration: N/A, bitrate: N/A
   Stream #0:0: Video: h264 (Baseline), yuv420p(tv, smpte170m, progressive), 768x432, 12 fps, 12 tbr, 1200k tbn, 24 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> wrapped_avframe (native))
Press [q] to stop, [?] for help
Output #0, null, to 'pipe:':
  Metadata:
    encoder         : Lavf58.29.100
    Stream #0:0: Video: wrapped_avframe, yuv420p, 768x432, q=2-31, 200 kb/s, 12 fps, 12 tbn, 12 tbc
    Metadata:
      encoder         : Lavc58.54.100 wrapped_avframe
frame=  596 fps=0.0 q=-0.0 Lsize=N/A time=00:00:49.58 bitrate=N/A speed= 179x
video:312kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
uxxxxx@s012-n001:~/oneAPI-samples/Libraries/oneVPL/hello-vpp/build$

Input file was extracted from mp4 file as follows:

ffmpeg -i ~/test.mp4 -c:v copy -bsf h264_mp4toannexb -f h264 test.h264

JaideepK_Intel · ‎08-20-2021

Hi,

Thank you for posting in Intel Forums,

We assumed that you were working on this sample(https://github.com/oneapi-src/oneAPI-samples/tree/master/Libraries/oneVPL/hello-vpp) So after going through the readme file, we could come to a conclusion that the sample you're trying to run doesn't support GPU. We've provided a screenshot for your reference. We will reproduce this soon and get back to you with updates.

Thanks,

Jaideep

Jitendra_S_Intel · ‎08-20-2021

Aside from using a compute node w/GPU, I didn't try to run the demo on the GPU unless the demo defaulted to GPU by using '-hw' option.

JaideepK_Intel · ‎08-24-2021

Hi,

We could reproduce your issue from our end & we observed that the hello_vpp & hello_decode samples are indeed giving wrong number of frames & different frame rate. We’ll investigate this issue further and get back to you.

And regarding the issue with “-hw option” i.e. Hardware Implementation, there are two types of implementations in Intel Media SDK & Intel OneVPL i.e. Software Implementation & Hardware Implementation. Software Implementation is used for running a demo sample on CPU. This can be done by adding “-sw” option in the command whereas the hardware Implementation is used for running a demo sample on GPU. This can be done by adding “-hw” option to your command. Since you added -hw option, you got the “no implementations” error. GPU compute node doesn’t effect this process.

Thanks,

Jaideep

Jitendra_S_Intel · ‎08-31-2021

Were you able to investigate this?

Jitendra_S_Intel · ‎08-24-2021

Where can I see supported target hardware for the '-hw' option? Application or documentation?

JaideepK_Intel · ‎08-25-2021

Hi,

You can check the supported hardware in the Readme.md file for every sample.

We've provided the link to the readme file(https://github.com/oneapi-src/oneAPI-samples/blob/master/Libraries/oneVPL/hello-vpp/README.md)

& below screenshot of it where the target device is mentioned as CPU.

That means that the sample is not GPU supported & doesn't work with "-hw" parameter.

Thanks,

Jaideep

JaideepK_Intel · ‎09-03-2021

Hi,

Sorry for the delayed response. It took us some time to figure out this issue.

Sample decode: please go through this readme file(https://github.com/oneapi-src/oneAPI-samples/tree/master/Libraries/oneVPL/hello-decode)

Currently sample-decode supports input format as h.265, since you gave input format as h.264 you got that error.

If you give h.265 file as an input and run the below command:

./hello-decode -sw -i <inputfile>(eg: test.h265)

The output will be stored as out.raw in the same directory (sample-decode gives output as I420 ) and to view this file run below command:

ffplay -f rawvideo -pixel_format yuv420p -video_size [width]x[height] out.raw

[width],[height] - width and height of the out.raw

Sample VPP: please go through this readme file(https://github.com/oneapi-src/oneAPI-samples/tree/master/Libraries/oneVPL/hello-vpp)

Sample vpp supports only I420/yuv420 as input colour format. Since we already got I420 file from sample-decode

now we can use that file as input to hello-vpp i.e

./hello-vpp -sw -i ../../hello-decode/build/out.raw -w <width of the out.raw> -h <height of the out.raw> -sw

Note: The output resolution for sample vpp will be 640x480(always)

The output will be stored as out.raw in the same directory (Output format:BGRA raw video elementary stream) and to view this file run below command:

ffplay -f rawvideo -pixel_format bgra -video_size 640x480 out.raw

Thanks,

jaideep

JaideepK_Intel · ‎09-06-2021

Hi,

If this resolves your issue, make sure to accept this as a solution. This would help others with similar issue.

Thank you,

Jaideep

JaideepK_Intel · ‎09-13-2021

Hi,

I assume that your issue is resolved. If you need any additional information, please post a new question as this thread will no longer be monitored by Intel.

Thanks,

Jaideep