Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6403 Discussions

failed to create engine: No OpenCL device found

ssava1
Beginner
‎04-09-2019 08:29 PM
1,951 Views

I am trying to run the openvino interactive_face_detection_demo I have setup a shell file to run it.  But the demo can not find the OpenCL device. 

First I checked the CPU

model name    : Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz

This CPU has a Intel® HD Graphics 4600. I have enabled it in the BIOS but it still does not show up. I think it is a BIOS issue.

Next I checked the VGA devices

:~$ lspci | grep VGA
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Pitcairn PRO [Radeon HD 7850 / R7 265 / R9 270 1024SP]
02:00.0 VGA compatible controller: NVIDIA Corporation GP107 [GeForce GTX 1050 Ti] (rev a1)

When I run the demo I get

InferenceEngine:
    API version ............ 1.4
    Build .................. 19154
[ INFO ] Parsing input parameters
[ INFO ] Reading input
[ INFO ] Loading plugin GPU

    API version ............ 1.5
    Build .................. 19154
    Description ....... clDNNPlugin
[ INFO ] Loading plugin CPU

    API version ............ 1.5
    Build .................. lnx_20181004
    Description ....... MKLDNNPlugin
[ INFO ] Loading network files for Face Detection
[ INFO ] Batch size is set to 1
[ INFO ] Checking Face Detection network inputs
[ INFO ] Checking Face Detection network outputs
[ INFO ] Loading Face Detection model to the GPU plugin
[ ERROR ] failed to create engine: No OpenCL device found which would match provided configuration:
    GeForce GTX 1050 Ti: invalid vendor type
    Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz: invalid device type

I understand why the 4771 can not be found, but The GTX 1050Ti is "invalid vendor type".  I am assuming that the clDNNPlugin is looking for Intel GPU.

To fix this issue, I think I need to do something with the plugin GPU. 

Yes, I know I can use CPU, but that is not want I want. The documentation is not good enough to fix this problem.  

Questions:

Is there a program to test the plugins?

How do I change  the GPU plugin and to what?

How do I find the source for the GPU plugin? I looked in https://github.com/opencv/dldt/tree/2018/inference-engine but can not find it GPU source files.

 

 

0 Kudos

Link Copied

4 Replies
nikos1
Valued Contributor I
‎04-09-2019 09:52 PM
1,951 Views

Sorry only Intel GPU devices for clDNN and also must be HD5xx or newer.

I believe you have an older HD4xx

There are similar threads in this forum - please search for clDNN or HD430

Cheers,

Nikos

0 Kudos
Copy link
ssava1
Beginner
‎04-09-2019 09:54 PM
1,951 Views

I have been doing more research on the problem. I found the program to test the OpenCl devices. https://gist.github.com/tzutalin/51a821f15a735024f16b

When I run It get.   Now this really confuses me. This program can find the OpenCL devices, why can not the openvino can not?

clDeviceQuery Starting...

2 OpenCL Platforms found

 CL_PLATFORM_NAME:     NVIDIA CUDA
 CL_PLATFORM_VERSION:     OpenCL 1.2 CUDA 10.0.292
OpenCL Device Info:

 1 devices found supporting OpenCL on: NVIDIA CUDA

 ----------------------------------
 Device GeForce GTX 1050 Ti
 ---------------------------------
  CL_DEVICE_NAME:             GeForce GTX 1050 Ti
  CL_DEVICE_VENDOR:             NVIDIA Corporation
  CL_DRIVER_VERSION:             410.104
  CL_DEVICE_TYPE:            CL_DEVICE_TYPE_GPU
  CL_DEVICE_MAX_COMPUTE_UNITS:        6
  CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS:    3
  CL_DEVICE_MAX_WORK_ITEM_SIZES:    1024 / 1024 / 64
  CL_DEVICE_MAX_WORK_GROUP_SIZE:    1024
  CL_DEVICE_MAX_CLOCK_FREQUENCY:    1392 MHz
  CL_DEVICE_ADDRESS_BITS:        64
  CL_DEVICE_MAX_MEM_ALLOC_SIZE:        1010 MByte
  CL_DEVICE_GLOBAL_MEM_SIZE:        4040 MByte
  CL_DEVICE_ERROR_CORRECTION_SUPPORT:    no
  CL_DEVICE_LOCAL_MEM_TYPE:        local
  CL_DEVICE_LOCAL_MEM_SIZE:        48 KByte
  CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE:    64 KByte
  CL_DEVICE_QUEUE_PROPERTIES:        CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE
  CL_DEVICE_QUEUE_PROPERTIES:        CL_QUEUE_PROFILING_ENABLE
  CL_DEVICE_IMAGE_SUPPORT:        1
  CL_DEVICE_MAX_READ_IMAGE_ARGS:    256
  CL_DEVICE_MAX_WRITE_IMAGE_ARGS:    16

  CL_DEVICE_IMAGE <dim>            2D_MAX_WIDTH     16384
                    2D_MAX_HEIGHT     32768
                    3D_MAX_WIDTH     16384
                    3D_MAX_HEIGHT     16384
                    3D_MAX_DEPTH     16384
  CL_DEVICE_PREFERRED_VECTOR_WIDTH_<t>    CHAR 1, SHORT 1, INT 1, FLOAT 1, DOUBLE 1


clDeviceQuery, Platform Name = NVIDIA CUDA, Platform Version = OpenCL 1.2 CUDA 10.0.292, NumDevs = 1, Device = GeForce GTX 1050 Ti
 CL_PLATFORM_NAME:     Intel(R) CPU Runtime for OpenCL(TM) Applications
 CL_PLATFORM_VERSION:     OpenCL 2.1 LINUX
OpenCL Device Info:

 1 devices found supporting OpenCL on: Intel(R) CPU Runtime for OpenCL(TM) Applications

 ----------------------------------
 Device Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz
 ---------------------------------
  CL_DEVICE_NAME:             Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz
  CL_DEVICE_VENDOR:             Intel(R) Corporation
  CL_DRIVER_VERSION:             18.1.0.0920
  CL_DEVICE_TYPE:            CL_DEVICE_TYPE_CPU
  CL_DEVICE_MAX_COMPUTE_UNITS:        8
  CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS:    3
  CL_DEVICE_MAX_WORK_ITEM_SIZES:    8192 / 8192 / 8192
  CL_DEVICE_MAX_WORK_GROUP_SIZE:    8192
  CL_DEVICE_MAX_CLOCK_FREQUENCY:    3500 MHz
  CL_DEVICE_ADDRESS_BITS:        64
  CL_DEVICE_MAX_MEM_ALLOC_SIZE:        3978 MByte
  CL_DEVICE_GLOBAL_MEM_SIZE:        15914 MByte
  CL_DEVICE_ERROR_CORRECTION_SUPPORT:    no
  CL_DEVICE_LOCAL_MEM_TYPE:        global
  CL_DEVICE_LOCAL_MEM_SIZE:        32 KByte
  CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE:    128 KByte
  CL_DEVICE_QUEUE_PROPERTIES:        CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE
  CL_DEVICE_QUEUE_PROPERTIES:        CL_QUEUE_PROFILING_ENABLE
  CL_DEVICE_IMAGE_SUPPORT:        1
  CL_DEVICE_MAX_READ_IMAGE_ARGS:    480
  CL_DEVICE_MAX_WRITE_IMAGE_ARGS:    480

  CL_DEVICE_IMAGE <dim>            2D_MAX_WIDTH     16384
                    2D_MAX_HEIGHT     16384
                    3D_MAX_WIDTH     2048
                    3D_MAX_HEIGHT     2048
                    3D_MAX_DEPTH     2048
  CL_DEVICE_PREFERRED_VECTOR_WIDTH_<t>    CHAR 1, SHORT 1, INT 1, FLOAT 1, DOUBLE 1


clDeviceQuery, Platform Name = NVIDIA CUDA, Platform Version = OpenCL 1.2 CUDA 10.0.292, NumDevs = 1, Device = GeForce GTX 1050 Ti
Intel(R) CPU Runtime for OpenCL(TM) Applications, Platform Version = OpenCL 2.1 LINUX, NumDevs = 1, Device = Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz

System Info:

 Local Time/Date =  04:49:07, 04/10/2019
 CPU Name: Intel(R) Core(TM) i7-4771 CPU @ 3.50GHz
 # of CPU processors: 8
 Linux version 4.15.0-47-generic (buildd@lgw01-amd64-001) (gcc version 7.3.0 (Ubuntu 7.3.0-16ubuntu3)) #50-Ubuntu SMP Wed Mar 13 10:44:52 UTC 2019


TEST PASSED

 

0 Kudos
Copy link
ssava1
Beginner
‎04-09-2019 10:10 PM
1,951 Views

Poor people can not afford new computers. I guess that limits development.

0 Kudos
Copy link
nikos1
Valued Contributor I
‎04-09-2019 10:44 PM
1,951 Views

I totally understand shaun s. 

Please read https://software.intel.com/comment/1931157 for more details why. We spent many days trying to get this to work but at the end was not worth supporting older GPUs that do not support Out-of-order execution

Please check clinfo :

  Queue properties
    Out-of-order execution             No
    Profiling                          Yes

Cheers,

Nikos

0 Kudos
Copy link
Reply

Community support is provided during standard business hours (Monday to Friday 7AM - 5PM PST). Other contact methods are available here.

Intel does not verify all solutions, including but not limited to any file transfers that may appear in this community. Accordingly, Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

For more complete information about compiler optimizations, see our Optimization Notice.