I won't have time to look at it, but Windows 10 Game Mode will be available in an insider preview released sometime today.  Apparently, it pushes background tasks to a couple of cores and leaving the bulk of the cores 100% dedicated to the game.  Reports are a 3 to 5% framerate improvement and smoothing.   Sounds very promising.   

 

Techradar gives info on how to sign up for insider previews.  Keep in mind if you install a preview, to go back you may have to completely reinstall Windows.  Game Mode will not be in general release until sometime in spring.

http://www.techradar.com/news/windows-10-game-mode-6-of-your-burning-questions-answered

 

https://www.rockpapershotgun.com/2017/01/25/windows-10-game-mode-explained/

http://www.windowscentral.com/windows-10-game-mode

http://www.theverge.com/2017/1/25/14390990/microsoft-windows-10-game-mode-explanation

 

 

Update:  build 15019 will not be out until later this week ... OK, it is later this week, so maybe today or tomorrow.  Anyway it did not arrive on Thursday as expected:

https://www.neowin.net/news/the-windows-10-insider-preview-with-game-mode-will-not-arrive-tomorrow

I don't really know, but, a lot (well, a few, anyway) of publishers have done their best to do this with windows as it is now, concerning their programs. It could be a monkey wrench for some of them.

I don't really know, but, a lot (well, a few, anyway) of publishers have done their best to do this with windows as it is now, concerning their programs. It could be a monkey wrench for some of them.

I've tried a lot to do things like move everything else to other cores and change priorities with Process Lasso with little to no effect, if any it maybe was placebo.  But this is Microsoft that is doing it in this case.  And they aren't just talking about CPU, but GPU priorities, how they're doing that I don't know, but it is their OS.  

 

Insider Update 15019 is finally on the download now.  Although it seems stuck at 0%, hrrmnn.  Update 15014 went smoothly after I turned off antivirus yesterday.  ...  waiting.

I've yammered a few times here how Windows fiber management technique is one of the least efficient methods compared to other OS.   And that some of the claims they made about DirectX 12 would be accomplished by fixing it.  I had also heard from somewhere that earlier DirectX API calls could be routed through DirectX 12 to improve their performance as well.  This may actually Finally be an implementation of that to some degree.    

The update is initializing now.

Thanks for reporting, Denali, much acknowledged.

 

Kind regards, Michael

I'm still not able to get 15019 to install.  Working the issue.

FYI, here is the official blog post about this build: https://blogs.windows.com/windowsexperience/2017/01/27/announcing-windows-10-insider-preview-build-15019-pc/#XugeEwCfjepeDb8E.97 

Because of the caveats listed there, I may not try too hard to get this build installed.  If it goes in I'll test.  But if not it may not be worth it anyways yet.

I finally got build 15019 to install.  Most of the trick was to be patient.   I do think I am seeing some improvement, even more smoothing!  And they're not yet done with implementing all the game mode improvements.   Here's what someone else saw.

 

I wonder if this would have any bearing on the functionality of the LVLD 767 in windows 10? Sure wish I could get that one flying again.

Something interesting.  I'm now on Build 15031.rs2_release.170204-1546.   So one thing that seems different is I can no longer use the windows key to bring up the task bar and bring up another running program on top.  I have to use alt-tab.  Knowing what I know about the event cascade in windows I'm wondering if they're cutting some of it out for a "Game" in game mode.   I really can't tell because I have once again borked my P3d, or something is borked, and I'm getting terrible framerates, 12-15 on a 5760 when I can usually get around 30.

 

But another interesting thing, is this:  I get the "close program to prevent information loss" with prepar3d.exe as the listed program, with the choice of close program or cancel.  Or I can just leave it alone in the background, and prepar3d keeps running no problem.  I don't know if this is the OOM that everyone sees, but when i look for that error message that's not the one everyone seems to be talking about with OOM.  I've had them before and in this case there is no bell in the background that just keeps ringing.  In fact I have prepar3d paused now after seeing and cancelling that message.  So is this something new, a new way Winblows is handling OOM for software like Prepar3d?  Like I said, I've been torture testing prepar3d and messed up something and I cannot tell anything about performance until I put it back together and close the hood, and I don't know if I've caused an OOM, or if this is something new.

 

Disregard.  I just started hearing bells, and then the crash.  Man, I haven't seen an OOM in quite a while.  What did I break?

 

As for this version of Windows preview, I am having issues outside of prepar3d, but they may be driver issues not related to prepa3d.  Or it could be the issue is affecting prepar3d, or not.  I'm just saying do your research before you jump in.  But there may be something here to work with. At any rate, this version is expected to be pushed out to everyone come April.

3-5%? Lol - like half a fps in P3D

 

 

So is this something new, a new way Winblows is handling OOM for software like Prepar3d?

 

It's possible, but I'm pretty sure the OOM dialog box that one sees presently is generated by P3d and not the OS. It happens when P3d requests a segment of VAS from the operating system that is larger than what is available. Other legacy apps like 32 bit Autocad and some of the 32 bit Adobe CS suite displayed different dialog boxes when running out of VAS.

A thought.  I've been making some photoscenery with FSEarthTiles, and they are an extra layer over other photoscenery I've made years earlier, so I can see what it looks like (in the SF area near SFO).  I'm just refreshing my photoscenery library with current sat photos.  I am wondering if the double layers of 30cm photoscenery in my scenery.cfg is causing OOM.

It's possible, but I'm pretty sure the OOM dialog box that one sees presently is generated by P3d and not the OS. It happens when P3d requests a segment of VAS from the operating system that is larger than what is available. Other legacy apps like 32 bit Autocad and some of the 32 bit Adobe CS suite displayed different dialog boxes when running out of VAS.

OOMs are not errors generated by a program.  They are unhandled exceptions/errors.  Windows is notifying the user that memory is not being managed by the program.  Windows could handle it better, and I thought maybe it was going to start.  But I guess not.

 

Sorry about the false alarm.

Here I am using photoscenery + custom AG with a triple view setup and never ecperienced an OOM..