NUC6i7KYK with Display Port 1.2 (MST)

[idata]

I have a NUC6i7KYK (Skull Canyon) device and it seems like the latest (and not 'greatest') drivers/firmware have taken a huge step backwards.

Basically, the device is not usable with the configuration I need (and is 'less' than what it is advertised' to support).

Machine is running Windows 10. The video's and info file is from the Insider Preview, but the problem occurs on 'Plain Vanilla' Windows 10.

Setup

HDMI : Dell U3014 (2560x1600)

DP : Dell U3014 (2560x1600) -> (MST) Dell U3014 (2560x1600)

Problems:

1. The display on the HDMI port is 'unstable'. Various drawing artifacts occur during composition. (Smooth background fade is the easiest to observe.)
2. Support for ANY of the monitors off of the DP connectors result in 1-3 second 'device resets' for things like moving the mouse on the display.

These problems occurred in nearly the same way with my NUC5i7RYH when it first arrived. (Subsequent firmware fixes and driver updates have made this machine usable.)

For comparison this is similar actions using the NUC5i7RYH:

[idata]

One thing I noticed that is suspect (at least to me) is that the drivers 'insist' that the monitor on HDMI is actually DP. The Intel Graphics utility shows the HDMI monitor as part of the the Display Port topology on the NUC6i7KYK and not on the NUC5iRYH. This at least COULD explain the behavior shown in the videos.

* Digital Television *

Display Type: Digital

DDC2 Protocol: Supported

Gamma: 2.2

Connector Type: DisplayPort

Device Type: Digital Television

[idata]

Hello, ErnieE:

Thank you for contacting the Intel Communities.Yes that configuration should be doable, due to the available bandwidth on the Intel® NUC Kit NUC6i7KYK.Let me further investigate this in order to better troubleshoot the scenario.All the information provided is really helpful, really appreciate it.I will keep you posted.Regards,Esteban C

[AHadd1]

Edited for Clarity : Issues I have are likewise on the NUC6I7KYK - Firmware 0034 - Graphics - 15.4444

Hi, I was going to start my own topic but this is pretty much spot on for me too though I'm using the DP connection (rather than HDMI) at 3840x2160.

I also had exactly the same issues on initial release of my older NUC5i5RYH which drivers fixed. Display can be perfect but sometimes on resume from sleep, monitor power off or even a full reboot I get the artefacts and occasional full screen flicker off and on.

It seems to be resolved by any of the below actions:

a) turning monitor off and on

b) Unplugging and re-plugging the display port cable

c) Changing the display resolution down and then back up again (easiest option)

I guess all this are forcing some form of displayport re-sync. I know its not the monitor or cable due to my prior experience with the NUC5i5RYH which was also used with the same setup and was fine post updates.

Cheers, Alex

[Ronny_G_Intel]

Hello,

I am sorry to hear that you are having issues, I can tell you that we are looking into this issue and I will get back to you as soon as possible.

Regards,

Ronny G

[idata]

It is frustrating to have had to wait so long for a 'Real' i7 NUC to arrive and then have it mostly unusable. I was exceptionally disappointed in the 'fake' i7 in the NUC5iRYH. I was also disappointed that Intel doesn't 'support' these devices using Linux. (For which that particular NUC was intended to be a Linux front end.) AFAIK the 'i7' in the NUC5iRYH was the first dual core i7. This was a complete shock as I would NOT have purchased that unit had I actually KNOWN that. The irony is that even with the device ~barely~ able to keep up with some of my LOCAL workloads, it is considerably more usable than a machine that is (as a whole) over double the performance.

I want to move to using the NUC6i7KYK as my primary 'thin client' on the setup shown, but given the (lack of) quality in the Firmware/Display Drivers (supporting MY setup) I can not. So far the investment has been heavy with very little return. The dual M.2 NVMe support works decent with Samsung 950's and having double the memory is nice, but having the screens flash on and off constantly + the corrupted output from the HDMI display makes the device useless. I really don't understand how such a step backwards was taken. Unless the support for the IGP was forked prior to the fix and no one bothered to integrate it, I can't fathom why Intel would release such a compromised (for ~MY~ setup) set of firmware/drivers. Given that I PLAN on using the NUC6i7KYK as a test for dual 4K displays this doesn't bode well.

I'd consider getting a 3.1c to DP adapter, but given the current support I don't really want to spend MORE money for something that should ALREADY WORK.

[AHadd1]

Not sure if you've seen but seem to be some new drivers and BIOS posted. 35 BIOS does't call out anything on the release notes but will install along with the 4454 Driver (I'm currently on 34 and 4444) and report back after a few days if see it fixed at my end.

[idata]

ErnieE, as per alexhaddock's indication I would like to invite you to test the BIOS 0035 and the graphics driver 4454 to check the behavior of your system.

In regards to the investigation on this matter, it is being handled for sure as rguevara commented above, but let me get an update from the investigation team in order to provide you with some information about that.I will keep you posted on this.Regards,Esteban C

[idata]

This morning I updated the firmware and video drivers and certainly have different behavior. A HUGE step backwards! After POST if I enter the firmware all three displays show the Intel firmware.

Now at the OS level NEITHER of the Display Port monitors works. The drivers DETECT them, but refuse to USE them.

I had to run into the office for a meeting so I didn't get a chance to get screen shots or videos.

I'll do that tonight or over the weekend.

Basically, only the single HDMI is supported at the moment.

[AHadd1]

That's a shame, I've had a massive improvement on mine since the update.

Updated BIOS to 35 - Still had issue (but the release notes didn't mention any graphics improvements to be honest)

Updated Chipset drivers 10.1.1.18 - rebooted

Updated Graphics to 4454 (it actually went to a black screen with cursor) - rebooted

Updated Graphics again (to be sure and it went fine that time) - rebooted

I also installed the newer ME drivers but doubt they have any relevance.

So far I've only noticed a couple of flickers (so perhaps a little more work needed) but nothing like the issues before.

I hope they fix your problem given how closely this mirrors the initial experience on the last NUC.

[idata]

The current display configuration is:

Display (L2R) Dell U3014 (DP via MST) | Dell U3014 (DP) | Dell U3014 (HDMI)

Cable DP to Mini DP | Mini DP to DP | HDMI 1.4

At the OS level I can NOT get the new drivers (20.19.15.4454) to recognize monitors via display port. I haven't tried disconnecting the HDMI display. I have tried various features of the firmware to select how displays are reported to the OS. This hasn't been an exhaustive test with all permutations as I have limited time to 'mess with this' before having to revert back to a working system. (Same video cables, keyboard etc works.)

NUC6i7KYK Attempt to extend desktop to DP

NUC6i7KYK Boot Ignoring DP at OS level

NUC6i7KYK Attempt to use SINGLE DP display

NUC6i7KYK at firmware setup

[idata]

Interestingly or annoyingly (depends on your point of view) after switching the cables back to the NUC5i7RYH and letting it boot with the exact configuration, I switched the cables BACK to the NUC6i7KYK. When the NUC6i7KYK came up, it ~finally~ decided the DP monitors were 'worth talking to', however, the behavior of the system has NOT improved AT ALL. The DP monitors wink on and off constantly and the HDMI monitor has the same artifacts shown in the original videos.

During the weekend my whole 'rig' was moved and the displays remained unplugged for at least 24 hours. When I did the firmware update on Friday, I had the displays setup so that I could selectively use three separate computers. (Which will eventually be duplicated in the 'new' location.) The NUC6i7KYK was the setup that got the same cable treatment. I don't know if the display driver is caching EDID information and this is causing issues. I have tried uninstalling and re-installing drivers.

Basically after hooking up the older NUC and then swapping the NUC6i7KYK back in, I am now back to the behavior in the original email. Sigh.

[idata]

Hi,I'm using the DP connection (rather than HDMI) at 3840x2160. I was going to start my own topic but this is pretty much spot on for me too though.

Thanks

http://www.faqtory.co/sky/ Sky

[idata]

Hello, mariakatosvich:

Thank you for letting us know about this from DP as well, I will be adding that to the investigation being performed.Regards,Esteban C

[YFeng6]

I have the same problem.

Acer K272HUL (2560x1440) with HDMI cable

Lastest BIOS Version

Windows 10 appears fine but BIOS keeps flashing.

[idata]

Thank you for your input on thisYu-Hao, the issue is being reviewed at the moment, as soon as I get a resolution on this I will let you know for sure.

Regards,Esteban C

[idata]

Hello, ErnieE:

Thank you for your patience on this, I just got an outcome from the investigation performed, it would be to install the HDMI 2.0 Firmware for your unit.Here is the link to download it: https://downloadcenter.intel.com/download/26061/HDMI-2-0-Firmware-Update https://downloadcenter.intel.com/download/26061/HDMI-2-0-Firmware-UpdatePlease let me know how your system behaves after installing this Firmware.Regards,Esteban C

[idata]

I am running firmware Version 35. I have attempted to install the HDMI Firmware update as a user and as an administrator. Attached to the NUC is a single HDMI 1.4 cable plugged into the HDMI port and on the other end is a Dell U3014 display using the HDMI port. The cable works fine on my other NUC, but just to be extra redundant I tried another cable and got identical results.

[idata]

I had a DP to HDMI cable that I could try. Same results attempting to update HDMI Firmware.

Dual HDMI is not quiet as flaky as DP | MST DP | HDMI, but I still get flicker. The very start of the attached video shows some, and at about 35-40 seconds shows more during the background switch.

Attachment -3.txt contains the info for the DP to HDMI cable.

Attachment -4.txt contains the info for DUAL HDMI (one via DP).

[idata]

AFTER RTFM (pdf) and not glossing over the implications of

NOTE: There can only be one video device plugged into the HDMI 2.0 port and it MUST be HDMI 2.0. Nothing can be plugged into the mini DisplayPort or Thunderbolt™ 3 ports.

I can 'forgive' the program for not working. I have NO HDMI 2 devices or cables (at the moment) to test this against. It is a little of a chicken and egg situation. I ~want~ to get a couple of UHD 4K displays and use them with this setup, but given the current poor showing, I have zero desire to sink additional money into this platform until it at least works for a setup that I have used going on 3 years. (Triple U3014s.) There is that, and the issue of > 32" displays is a bit of over kill for this machine. (Unless curved or IPS, but after using a 40" UHD 4k display @ 30hz for 3-4 months, I know that it is WAY to large for ~MY~ usage. 30Hz is 'livable', but I've been holding out for HDMI2 and 60hz.

I am also dreaming for something along the lines of a 'single dual curved OLED UHD 4K' display. (a curved dual display with NO bezel would be super cool). AFAIK no such device exists. I wouldn't mind if the display required inputs for each 'half' of the 4K. ( 7680x2160) The current 34" WQHD displays have way too small of a vertical resolution (ex. 34UC88-B 34, 3440x1440). I've been dual display from around 1997 and triple (16:10, all Dell, 24" -> 27" -> 30") for almost 10 years. I am NOT a fan of vertical displays for ME. I've tried 1600x2560 | 2560x1600 | 2560x1600 | 1600x2560 (quad U3014), but ultimately I want something uniform.

Make a 30-32" (curved or not) 4K (3840x2160) OLED display and I would be VERY temped to upgrade again. LG has some extraordinary 4K OLED (super thin. tiny bezel) displays, but they have (at least to date) been 'smart TV' style and HUGE without HDMI2 (http://www.lg.com/us/tvs/lg-65EF9500-oled-4k-tv LG 65EF9500: 65 Class (64.5 Diagonal) Flat OLED 4K Smart TV w/ webOS 2.0 | LG USA ) this 'TV' is beautiful. I'd also love to see something like a continuous curved 7680(+)x2560(+) display. The 'panel fit thing' from Dell would come in handy then. Last thing ~I~ want is a maximized editor across the entire display!

Blah, Blah, Blah.

Gotta got back to the working NUC now. Still massively disappointed in a 'flagship' device for MY setup. (A single 2560x1600 HDMI display is stable, but even ~just~ two HDMI displays is painful.) So far the $$$ that I have sunk into this device has produced negative ROI. Still massively disappointed.

--Ernie

.

[idata]

Thank you for providing all the proves and testing performed in your unit ErnieE, and letting us know about the setups you would like to get, they sound great indeed (personal opinion).

This information is certaily helpful to assist us in the fix for this situation, I will be using this data to further investigate the scenario and the provide you with more troubleshooting to be performed or the fix.Will keep you updated on this matter.Regards,Esteban C