MSDK cant support HW acceleration if system connect 2 displays

[Zheng_L_]

my notebook computer has a NV MX150 independent card and a intel UHD620 integrated card,my program use nv card to do CUDA computing and intel card to do hw accelerated video decoding/encoding, my program works fine when system only use notebook's display,but when I use a second display connected to notebook's HDMI port,the MSDK HW accelerated video decoding/encoding can not work,MFXInit() return MFX_ERR_UNSUPPORTED; I found I must bind my program to Intel card,then the HW acceleration can be worked,but that my program can not do the cuda computing;  nv card,msdk acceleated dec/enc,second display are all my program needed,so how can I get it?

[Ramashankar]

Hi Zheng,

I believe you need not to bind your program to any particular card. You can just try to init your session using MFXInit with various hardware flags.

Refer this post: https://software.intel.com/en-us/forums/intel-media-sdk/topic/780596#comment-1923855

Hope it helps.

[Zheng_L_]

Ramashankar wrote:

Hi Zheng,

I believe you need not to bind your program to any particular card. You can just try to init your session using MFXInit with various hardware flags.

Refer this post: https://software.intel.com/en-us/forums/intel-media-sdk/topic/780596#com...

Hope it helps.

Hi,Ramashankar:

   I use MFX_IMPL_HARDWARE_ANY to call MFXInit(),and I also try the MFX_IMPL_HARDWARE2/3/4,but all failed, but my program init the CUDA SDK firstly,then init msdk, am I going to reverse the order?

[Mark_L_Intel1]

Hi Zheng,

Did you check Appendix D of Media SDK developer guide?

Mark

Let me know if this solves your problem.