OP here OMG, what have I done, what have I unleashed!
So many pros and cons, and so many use cases.
First, thank you all for the help.It will take a wile to reply to so many advises and questions, so please don't get mad at me if I can not reply to each, as I should.
Briefly:
- tearing is happening with or without proprietary nVidia, with or without Vsync. In fact, Vsync is possible on only one monitor at once, and only in full screen. So, if not full screen, or if I move the movie to another monitor, will tear the video. So far I tried the nVidia blob (both stable and latest beta drivers) from nVidia website (with Ubuntu, manual install), and nVidia proprietary, but installed from the Fedora repository (with Fedora, install from GUI).
- about tearing again, the nVidia drivers (all 3 I've already tested) have two check boxes in the GUI for Composition Pipeline. There are boxes for each monitor, yet Composite Pipeline can fix tearing on one monitor at a time, the one driving the Vsync. Tear will still happen on another monitor. Need to look further into this, maybe if I write the parameters in the config file, than it will work with all of the monitors at once. Mainly I tested with VLC, but YouTube videos (Firefox) seems to be even worst, and ignore any settings.
- about tearing, the last thing to say, I noticed when there is only one monitor enabled, and the video is full screen, then the BLIT mode written in red by the driver changes to FLIP mode written in green. This watermark with the video mode is slapped on the video by the driver, and it does not appear in a full screen YouTube video, yet the driver video mode stamp appeared in a very unexpected place like the Microsoft's Visual Studio IDE (under Linux, installed from repo's GUI, not Wine). I read Visual Studio was running like a webpage, but I don't really know. No idea why an IDE like Visual Studio running in a window will invoke a video mode from the nVidia driver (thus the driver watermark stamp), but a YouTube played in full screen with Firefox will not.
- about repository of each Fedora and Ubuntu, I understand there is not much difference between the packages in each repo, which is good news. In practice, there seem to be some stability issues, didn't dig into that yet. An example: If I install Meld from each distro (Meld == a GUI interface for the GNU diff tool) in order to diff a huge pile of files (I am in the middle of saving and recovering parts of personal data from the former C:\ of the Win10) the Ubuntu Meld crashes, the Fedora Meld works.
- again, I am not new to Linux and loved it since I can remember. Was using Linux over the years at work, but only in embedded or industrial environments where a SSH terminal was more than enough. No one cares about sound, or video, or stanby/power features for a headless industrial device. Yet, I didn't used Linux as a day to day (multimedia) desktop at home. Periodically, let's say at each 3-5 years, I am giving Linux a try as a home desktop, in order to get rid of Windows.
- never completely switched to Linux until now. Maybe because of occasional game playing, and also because of the daunting task of cleaning/saving personal data that piled up over the years, then migrate all kind of settings, setups and habits.
Let me give some examples: There are 2 stereo speakers (with amplif) and a headset connected to the onboard 7.1 sound card. They are all connected, all the time, and I was redirecting the sound from one to another from software (under Windows, the Creative sound drivers comes with their own gui that can redirect the sound). There was no need to unplug the physical jack connectors. With Linux, the default sound driver doesn't have such an option from GUI. I need to dig for new drivers or customized settings, or find some workaround. For now, I don't even know if the sound server is with PulseAudio or something else.
Same with video: There are 3 monitors connected at nVidia, and a 4'th one connected at the onboard Intel GPU with a DVI cable that is going to the other room, for the movies only. For movies on the 4'th monitor, there is another wireless keyboard and mouse, in the same room with the 4'th monitor. Wireless keyb and mouse still works under Linux, yet the keyboard's mutimedia buttons for volume UP/DOWN doesn't work. Need to tinker with that one, too.
Another one with the wireless sound: There is a CSMR wireless audio and a wireless headset that I use to listen while doing something else around the house. The wireless sound doesn't works either. It doesn't detect the CSMR transmitter at all. Need to tinker with that one too, or live without it.
Small commodities like that piles up, and take a lot of time to set them properly.
TL;DR again1. I'm not going back to Windows 10, no matter what. WinXP was great. Win10 is horrible, and yes, it made my blood boil way too often. The trend with Win10 is not for me, no reasons to look back. I'm almost thankful to that malware I got last year. It made me switch to Linux at home.
2. Since I'm not in a hurry, and don't need to commit to a certain distro, I decided to go for plan B:
Level1 hypervisor + ZFS filesystem + hardware passthrough It will be a pity to have a processor that supports virtualization and not use it. I'll go ahead and aim for a personal cloud, with distributed storage and multiple VM. Will probably go for ProxMox, which is KVM based but comes nicely bundled with ZFS file system and WebGUI for administration. ProxMox is the "ubuntu of level1 hypervisors". I tried Xen once, and needed to tinker a lot to make it work, and there was no free GUI for Xen at the time. I also read AmazonAWS plans are to switch from Xen to KVM, too.
3. Can hardly wait to tinker with hardware pass-through, GeForce760 might need a hardware unlock to allow hardware pass-through. If got it right, the the chipset can do that, GTX760 is a Quadro locked down, just waiting for a hack. The GTX760 unlock will also suppose to enable the 10bits/each color over HDMI (30 bits color depth is another Quadro feature locked down in the low end GTX760). The main monitor can do 10bits/4K/60Hz.
I am very curious if the difference between 24 and 30 bits in color depth is visible to the eye, or it is just a marketing thing.