I don't have a system to check immediately. I have asked some developers for a break down. Current expectation is that between 6th gen processors codename skylake and 7th gen processors codename kabylake there isn't anything that should preclude functionality.
EDIT: Those SKUs are not currently supported. The driver itself is expected to perform device id checks against the known hardware listed. User mode runtimes can be linked and built on a wider ranging set of platforms... but for execution... the platforms laid out here are the only ones expected to function: https://github.com/intel/media-driver/. Please reference this form post: https://software.intel.com/en-us/forums/intel-media-sdk/topic/735307
Have you tried building an application against the latest checkin here: https://github.com/Intel-Media-SDK/MediaSDK? Did you experience issues? Can you share what they were if you did.
I'm confused with following descriptions:
From https://github.com/intel/media-driver/. The supported platform
From https://github.com/intel/intel-vaapi-driver/releases, Coffee lake is supported
BTW,I will buy I3 8100(CFL) and Z370 board to test.
If you are using the guide to build the open source media stack(https://software.intel.com/en-us/articles/build-and-debug-open-source-media-stack), you have to refer to the release note of https://github.com/intel/media-driver because this is the driver the open source Media SDK is using.
During the build process by the guide, the default media driver will be replaced. This built media driver was validated by the Media SDK team and it should have the good quality.
I confirm that media driver(open source Media SDK software stack) not support I3 8100(coffeelake).
We must wait cannonlake for about one year.
My question is if MSS support cannonlake in the future? Should we move our project from MSS to media driver for future cannonlake?
This is a good news and let me if you have further questions.
Under current status, I recommend to use master to build MSDK library when you refer the article I mentioned in my previous post. The tested commit ID in the article has some work around and also has an issue which causes rendering problematic.