Support NeoGAF

[MaLDo]

Could be. In my case stutters are always in the same point. Only three people in the thread can see them and in the same moments so maybe something is fucked in the conversion. Will download the video from YouTube and try it. Interesting if YouTube can add in this FH3 video the exact same problem than the game has.

Maybe you should step back and consider that the problem is with YouTube, and not the game or the video that Microsoft recorded?

It's really not hard to understand here. No one is lying to you. People - including me - are not observing the same stutters that you are.

YouTube automatically re-renders videos in countless different formats for different devices and different browsers which support different things. Maybe one of them got screwed up?

I personally am also not seeing any stutters using Chrome on Windows 10.

After some test the problem is in the version of the video in YouTube server that iOS uses.

Tested in iphone 5, 6 and 7. All three show the stutters in the same time every time. I recorded in slow motion the video (iPhone 7):

First stutter
https://m.youtube.com/watch?v=8hBOtdE94ho

Second
https://m.youtube.com/watch?v=7G19GJior70

Third
https://m.youtube.com/watch?v=lPNZJ2T5Fu8

 

[Trup1aya]

Windows Game Mode delivers frame rate boosts of 2 to 5pc

http://hexus.net/tech/news/software/101881-windows-game-mode-delivers-frame-rate-boosts-2-5pc/

That´s some serious boost, don´t you think?

I wouldn't say that's "serious" but the frame rate improvements are going to vary depending on how Much CPU would have been bogged down, had gamemode not been enable.

If this particular 2-5% gain is in a scenario where gamemode wasnt even contending with any unnecessary tasks, I'd say 2-5% is pretty good.

If there were a bunch of highly taxing background tasks getting affinitized, I'd have expected bigger gains.

Whats more important than average gains, imo is something that prevents dips and spikes whenever random background tasks try to ramp up.

 

[Lashley]

Any improvement is an improvement

 

[razorblade79]

Will game mode unlock the secret 60fps version of FH3?

I hope not, I'd be sad if FH3 would only run with 60fps

 

[dr_rus]

It works on any game. Both UWP and Win32. Seriously. It's right there in the OP.

The figure of 2-5% boost given for UWP titles only which are "getting the most" of this mode. So if it's borderline unnoticeable for UWP then you can guess what it will be for Win32.

And as for frametime stability - it is very unlikely that removing the hitches caused by background processes will mean a difference between a smooth 30/60 and 30/60 with constant hitches. A properly configured Windows running on a quad core PC does not cause often hitches in games, it's more of an exclusion from the rule when this is happening. So I'm not really expecting much on this side either.

 

[RazorbackDB]

The build including game mode is up, but I really don't feel like getting back on the fast ring right now.

https://blogs.windows.com/windowsex...r-preview-build-15019-pc/#crmEQfMx2CC6ZXKe.97

 

[Zedox]

The build including game mode is up, but I really don't feel like getting back on the fast ring right now.

https://blogs.windows.com/windowsex...r-preview-build-15019-pc/#crmEQfMx2CC6ZXKe.97

Yea, I was going to put that link in here. I'm not sure if we should make a thread about it or not. It is a preview build with a bunch of problems.

 

[Neo_Geo]

I'm not liking the "Popular games such as may experience crashes or black screens when trying to load due to a platform issue." I was going to dive in and test for a bit, but I can almost guarantee that a new install would be needed sooner rather than later with the laundry list of issues.

 

[gamz]

It's out for PC now.

 

[Soren01]

Waiting for impressions.

 

[VeteranGamer]

I'm not planning on using their store much, but good for those who are.

Did you read? Steam too

 

[Ricerocket]

It's out for PC now.

So how is it? Any additional options other than on/off switch in the gamebar?

It really just sounds like exclusive fullscreen mode vs borderless window. If it's true that GameDVR is moving to settings and can be disabled for performance gains without an account, then that is a huge positive for me.

 

[dLMN8R]

Well, it's optional and disabled by default because things aren't fully tested yet for perfect compatibility in every game. So it's very limited for now and will roll out gradually as more games and hardware configurations are tested.

 

[Dinjoralo]

Fullscreen games work with the bar, but it's only available in a small pool of games. For testing, I'd assume, because trying to fix things people are having with the thousands of games on steam sounds like a nightmare.

The games are Battlefield 3, Call of Duty: Black Ops 2 + Zombies, FIFA 14, FIFA 17, FIFA Manager 14, Grim Dawn, Guild Wars 2, Left 4 Dead 2, MapleStory, Paragon, Payday 2, Rocket League, The Elder Scrolls Online, The Sims 4, Tom Clancy’s Rainbow Six Siege, and Warface.

I haven't been able to test it myself, but you can have the bar up for using game mode while also making it not record.

 

[Head.spawn]

After some test the problem is in the version of the video in YouTube server that iOS uses.

Tested in iphone 5, 6 and 7. All three show the stutters in the same time every time. I recorded in slow motion the video (iPhone 7):

First stutter
https://m.youtube.com/watch?v=8hBOtdE94ho

Second
https://m.youtube.com/watch?v=7G19GJior70

Third
https://m.youtube.com/watch?v=lPNZJ2T5Fu8

That's weird as hell. You made me think I was going crazy for a bit, I actually sat there counting frame by frame thinking I'm losing it, lol.

 

[Paragon]

So far I've tried using this with Resident Evil 7, The Witcher 3, Dishonored 1, and DOOM.
I had to switch them into windowed mode, enable the option via the game bar, restart the game, and then put them back into fullscreen mode for it to work.
In all of these games the option introduces a really bad stutter that happens every second or two. Even the mouse lags badly.
Has anyone else tried this and had different results?

Edit: Well whatever it was, it doesn't happen now.

 

[Iced_Eagle]

I tried GameMode using Tom Clancy's Rainbow Six Siege. I used the in-game benchmark utility. Ran it twice w/out Game Mode and grabbed the second result. Then enabled Game Mode, ran it twice and grabbed the second result.

Running Med-High settings @ 1440p on a 144Hz monitor. Using GSync as well.

i7 4770k
16GB RAM
Nvidia 780 Ti

(using screencap of my post on Reddit with the results since Gaf doesn't support tables?)

Top table is with GameMode disabled:

The overall average is essentially within the margin of error. No change there. It did raise my min FPS though by a small amount which is sorta nice, and my max fps had both positives and negatives.

 

[mike4001_]

Thanks, this actually looks pretty good for what was expected.

Minimum fps are better because no process can highjack the system for a second and lower your fps for that fraction (and in edge cases cause a stutter).

Strange that the average is lower but this might just be the margain of error.

 

[LelouchZero]

I tried GameMode using Tom Clancy's Rainbow Six Siege. I used the in-game benchmark utility. Ran it twice w/out Game Mode and grabbed the second result. Then enabled Game Mode, ran it twice and grabbed the second result.

Running Med-High settings @ 1440p on a 144Hz monitor. Using GSync as well.

(using screencap of my post on Reddit with the results since Gaf doesn't support tables?)

Top table is with GameMode disabled:

The overall average is essentially within the margin of error. No change there. It did raise my min FPS though by a small amount which is sorta nice, and my max fps had both positives and negatives.

Oh that's interesting, what hardware do you have btw?

 

[Iced_Eagle]

Oh that's interesting, what hardware do you have btw?

i7 4770k
16GB RAM
Nvidia 780 Ti

(Good question. I'll add that to my post).

Thanks, this actually looks pretty good for what was expected.

Minimum fps are better because no process can highjack the system for a second and lower your fps for that fraction (and in edge cases cause a stutter).

Strange that the average is lower but this might just be the margain of error.

Yeah, average certainly seems to be within a margin of error. Overall difference of 0.4 fps is pretty much squat.

*Edit* Decided to try The Division benchmark. Unfortunately their benchmark doesn't give a whole lot of detail, but it's pretty much identical results disable/enabled based on what they show.

Disabled:
Avg FPS: 51.9
Typical FPS: 52.2
Avg CPU: 43%
Avg GPU: 98%

Enabled:
Avg FPS: 51.9
Typical FPS: 52.1
Avg CPU: 42%
Avg GPU: 98%

 

[Sad Affleck]

I tried GameMode using Tom Clancy's Rainbow Six Siege. I used the in-game benchmark utility. Ran it twice w/out Game Mode and grabbed the second result. Then enabled Game Mode, ran it twice and grabbed the second result.

Running Med-High settings @ 1440p on a 144Hz monitor. Using GSync as well.

i7 4770k
16GB RAM
Nvidia 780 Ti

(using screencap of my post on Reddit with the results since Gaf doesn't support tables?)

Top table is with GameMode disabled:

The overall average is essentially within the margin of error. No change there. It did raise my min FPS though by a small amount which is sorta nice, and my max fps had both positives and negatives.

Very promising, I was hoping for an increase in minimum framerates. I hope we get some benchmarks for frametimes too, hopefully it helps there as well.

 

[StromboliProtocol]

Source: http://m.windowscentral.com/windows-10-game-mode



More details in the source, looks like a nice feature

The XBox game bar ???

Is it a typo? I thought it was about PC version of Windows 10...

 

[Theorry]

Iced_Eagle do you have Gears 4?

 

[omnislayer]

The XBox game bar ???

Is it a typo? I thought it was about PC version of Windows 10...

Windows 10 on PC has a thing called the Xbox game bar...

 

[orava]

I'll take consistency over higher frame rates anytime.

 

[Soren01]

Iced_Eagle, can you run the benchmark with vsync or fps limiter on and post the MSI Afterburner frametime graphs?

 

[dr_rus]

Guys, I hope you understand that you can have such change in minimum fps without enabling any "game mode", just by going through the same sequence for a second time.

 

[leeh]

Guys, I hope you understand that you can have such change in minimum fps without enabling any "game mode", just by going through the same sequence for a second time.

I'm sorry, but why are you so hellbent into putting a negative spin on a feature that's universal and optional? It's also currently in the fast ring, so hardly finished.

 

[lordfuzzybutt]

Guys, I hope you understand that you can have such change in minimum fps without enabling any "game mode", just by going through the same sequence for a second time.

So? He takes the 2nd run results for both Enabled and Disabled, and Enabled minimum result is higher in each and every scene. That's pretty consistent and obvious that it seems to work.

 

[Neo_Geo]

Need to do 3-5 benchmarks with it enabled/disabled and then average each set of results, I also dislike running two benchmarks and then taking the second run to determine the performance, not a very good indicator. Thanks for preliminary test results though, it does show promise so far.

EDIT: Not my results, someone did a quick test, Game Mode Off (L) Game Mode On (R)

http://www.3dmark.com/compare/fs/11546334/fs/11546356

 

[Pooya]

Some results on lower end CPUs like Core i3 and similar are more interesting to see, they would probably benefit more than stronger CPUs that have a lot of spare time anyway.

 

[Neo_Geo]

That is true, similar to the "BIG" Nvidia release that improved DX11 performance to near Mantle levels, it was only really beneficial to the low end CPU market. Modern Intel overclocked 4c|8t+ CPU's probably won't see much of an improvement unless the user has a badly configured machine with processes eating abnormal amounts of CPU cycles.

 

[Paragon]

I tried it again after a cold boot of my PC today and wasn't getting that severe stuttering and mouse lag that I had before.
I'm not sure what causes it, but it seems like it could be related to alt-tabbing away from fullscreen games and back to them.
Doesn't seem to happen every time though.

However I'm still getting occasional stutters in games which run perfectly smooth without using game mode, so my system doesn't seem to like whatever it's doing.

That is true, similar to the "BIG" Nvidia release that improved DX11 performance to near Mantle levels, it was only really beneficial to the low end CPU market. Modern Intel overclocked 4c|8t+ CPU's probably won't see much of an improvement unless the user has a badly configured machine with processes eating abnormal amounts of CPU cycles.

Every CPU benefited from the addition of driver command list support. You just have to look at minimum framerates and not averages. NVIDIA are still outperforming AMD by a large margin when you look at minimum framerates in many DX11 titles due to this.

 

[M3d10n]

Sounds like sort of a new process priority tier, similar/inspired to what they use on the Xbox One, but which also includes GPU resources in the mix.

Since you can open the game bar on pretty much every application, not only games, it should lead to some fun experiments, like enabling game mode on Photoshop.

 

[Iced_Eagle]

I can't take any requests for more tests unfortunately. Rolled back to the slow ring build last night. That build had pretty big issues with games such as Overwatch and Arma 3 not working at all. I filed my bugs and headed back to the old build but will try again with a future build.

I also tried out Beam broadcasting. Ridiculous simple to set up, but unfortunately it consistently crashed less than 5 minutes after starting a broadcast each time.

 

[LightningXCE]

New nVidia drivers seem to break a lot of things, including hardware encoding, FWIW.

 

[Theorry]

I can't take any requests for more tests unfortunately. Rolled back to the slow ring build last night. That build had pretty big issues with games such as Overwatch and Arma 3 not working at all. I filed my bugs and headed back to the old build but will try again with a future build.

I also tried out Beam broadcasting. Ridiculous simple to set up, but unfortunately it consistently crashed less than 5 minutes after starting a broadcast each time.

No problem dude. Was about to ask about Beam indeed. Sounds good. Except for the crashing offcourse.

 

[Iced_Eagle]

No problem dude. Was about to ask about Beam indeed. Sounds good. Except for the crashing offcourse.

Yeah, it was my first time ever trying streaming too! Was playing Rainbow 6, opened the game bar, and hit the broadcast button. It took around 20-30 seconds for it to get started (spinning circle thing in the game bar), then a purple overlay window showed up and gave me some options on Stream Name, webcam/mic stuff, whether to stream only the game window or everything, etc, but I just hit start and it then showed up on my Beam channel seconds later! The UI on my screen was minimalist. Basic viewer count and all that, and there was a large rectangle you could alternate between either a preview window and the chat. I didn't play with it enough to test, but I hope you can drag that window over to a second monitor. Even though it was minimalist, if you have a second monitor, it makes sense to be able to get that completely out of the way.

Latency was crazy good (which is Beam's strength after all). I had my channel opened in a browser on my second monitor. Seeing me move my mouse and like less than a second later it showing up on the channel was really cool.

Obviously the bugs and all that will be worked out, so whatever, that doesn't worry me. I filed a bug anyway, so next better build for gaming, I'll give it another shot. I do think this has a good chance to get more people doing streaming! It was literally three button presses to start (Open GameBar -> Broadcast -> Start).

 

[dr_rus]

I'm sorry, but why are you so hellbent into putting a negative spin on a feature that's universal and optional? It's also currently in the fast ring, so hardly finished.

Because I don't like it when MS is spending time on a feature which is of no benefit to the end user instead of making something really useful, like providing the OS level ability to navigate the desktop with a gamepad? Is that so hard to understand?

So? He takes the 2nd run results for both Enabled and Disabled, and Enabled minimum result is higher in each and every scene. That's pretty consistent and obvious that it seems to work.

Judging anything from minimal fps result of the whole benchmark is always misleading, if you want to look at the spikes you have to make a frametimes graph. One number does not provide enough information to make anything out of it.

 

[lordfuzzybutt]

Judging anything from minimal fps result of the whole benchmark is always misleading, if you want to look at the spikes you have to make a frametimes graph. One number does not provide enough information to make anything out of it.

No shit? Your original post contributed the higher minimum fps to running the sequence a 2nd time. Which I then argued that the results are already from the 2nd time. That's it. The whole argument is about the minimum FPS.

I am not arguing that it definitely works in overall, but that it seems to work in raising the minimum FPS. Like you said, we need a whole suite of benches with graphs before we can conclude if this feature is something useful or not.

 

[therealbigrich]

2-5% higher fps for free. Without doing anything, except clicking once. How the hell did this ever become a bad thing?

When it's released by MS anything can be a bad thing

 

[therealbigrich]

After some test the problem is in the version of the video in YouTube server that iOS uses.

Tested in iphone 5, 6 and 7. All three show the stutters in the same time every time. I recorded in slow motion the video (iPhone 7):

First stutter
https://m.youtube.com/watch?v=8hBOtdE94ho

Second
https://m.youtube.com/watch?v=7G19GJior70

Third
https://m.youtube.com/watch?v=lPNZJ2T5Fu8

Didn't notice any stutters in the video myself . Did you try looking at the video on a PC or other than iPhone at least . just curious

 

[the thoroughbred]

It always annoys me that an iPad or similar device with a much smaller battery can stream videos for hours and my laptop can barely last a movie. I thought clearly the resolution on my laptop is lower and the battery weighs more than the iPad itself. I realised that the computer is doing many other tasks in the background. Which I feel are totally unnecessary. I just want to stream this and the battery to last as long as possible. Now they've done this with games.

Sometimes I really wonder what the operating system is doing taking up so much power. They must be monitoring every keystroke, watching through the camera using the mic and sending it to GCHQ / NSA.

 

[LukasTaves]

Because I don't like it when MS is spending time on a feature which is of no benefit to the end user instead of making something really useful, like providing the OS level ability to navigate the desktop with a gamepad? Is that so hard to understand?

The feature came from the Xbox side, very likely because they are going to get rid of the second os and needed a single os that prioritized games while allowing apps to have full resources when the game was not running.

As for using gamepad to navigate the desktop then you are going to be happy soon.

Ms is changing the shell of Windows to a new one based on the same xaml uwp apps have access to. They want to reimplement all their devices interfaces with that new shell so for example you can have a phone that unfolds into a tablet, that can be docked to a monitor and show the desktop ui or docked to a tv and have a Xbox like interface.

And along else, there's an ongoing project of replacing every win32 part of the Windows (including the file explorer) to a uwp equivalent. This is good because uwp apps have native support for controller inputs, though the app is pretty barebones compared to the uwp one which is why I think they haven't released it yet.

 

[singhr1]

Because I don't like it when MS is spending time on a feature which is of no benefit to the end user instead of making something really useful, like providing the OS level ability to navigate the desktop with a gamepad? Is that so hard to understand?



Judging anything from minimal fps result of the whole benchmark is always misleading, if you want to look at the spikes you have to make a frametimes graph. One number does not provide enough information to make anything out of it.

Woah narcissist, relax. You act like that shit is easy. Everything they are doing is benefiting someone. Sometimes that someone isn't you.
Also, can't believe people are under the illusion that someone working on some feature is taking resources away from developing something else. Diff peeps for diff development teams.

 

[opricnik]

Because I don't like it when MS is spending time on a feature which is of no benefit to the end user instead of making something really useful, like providing the OS level ability to navigate the desktop with a gamepad? Is that so hard to understand?



Judging anything from minimal fps result of the whole benchmark is always misleading, if you want to look at the spikes you have to make a frametimes graph. One number does not provide enough information to make anything out of it.

Really? Imagine how will this be affective for laptop gamers stop being entitled

 

[Theorry]

Yeah, it was my first time ever trying streaming too! Was playing Rainbow 6, opened the game bar, and hit the broadcast button. It took around 20-30 seconds for it to get started (spinning circle thing in the game bar), then a purple overlay window showed up and gave me some options on Stream Name, webcam/mic stuff, whether to stream only the game window or everything, etc, but I just hit start and it then showed up on my Beam channel seconds later! The UI on my screen was minimalist. Basic viewer count and all that, and there was a large rectangle you could alternate between either a preview window and the chat. I didn't play with it enough to test, but I hope you can drag that window over to a second monitor. Even though it was minimalist, if you have a second monitor, it makes sense to be able to get that completely out of the way.

Latency was crazy good (which is Beam's strength after all). I had my channel opened in a browser on my second monitor. Seeing me move my mouse and like less than a second later it showing up on the channel was really cool.

Obviously the bugs and all that will be worked out, so whatever, that doesn't worry me. I filed a bug anyway, so next better build for gaming, I'll give it another shot. I do think this has a good chance to get more people doing streaming! It was literally three button presses to start (Open GameBar -> Broadcast -> Start).

Yeah sounds good. Just like Xbox really. Sure you are abit limited compared to setting it all up by yourself with a capture card, OBS etc. But this is still good quality and very easy to setup. So more will stream.

 

[GravyButt]

Will this help at all with any sort of streaming of games?

I have had 0 problems streaming on wired connections (it was like I was playing directly on my tv it was that good) but all of a sudden since my xbox did some update the other week (I think it was last week maybe?) streaming has been horrible and unusable. Hopefully this can help. Not sure what the heck happened all of a sudden.

 

[M3d10n]

Someone needs to test Dolphin on this.

 

[EvB]

When it's released by MS anything can be a bad thing

I can't tell if that is an sarcastic observation of GAF's general anti-Microsoft rhetoric or not.