FIBER_FRAME_TIME_FRACTION

[vgbaron 3,632]

TANSTAAFL - for those not familiar - There ain't no such thing as a free lunch.

 

The increase in FPS comes at the expense of something - as some have noted, certain autogen etc doesn't load. If you are fine with that, great - but that will be an individual call.

 

Originally FSX defaulted to .33 but with newer hardware we began moving that around. With P3D and them off loading stuff to the GPU we have more "wiggle room" than before.

 

Since I keep my system locked at 22 and it's smooth as silk my improvement will be minor but it appears that it can be a major help to many.

 

Vic

[GaryGB 10]

I wish I hadn't clicked this topic - you've got me tweaking the cfg which I promised myself I wouldn't do!

 

Will post results tonight.

 

You guys are on a roll !

 

Keep on peeking at P3D (...and FSX) tweaking ! 

 

 

 

Some more tweaks offering possible 'help' for P3D optimization by the end user:

 

https://polypoke.wordpress.com/2011/01/10/fsx-cfg-tweaks/

 

 

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CB4QFjAA&url=https%3A%2F%2F#####.wordpress.com%2Ffsx-oom-and-addon-vas-usage%2Fcomment-page-1%2F&ei=KvdcVJDVLNGdygScyIHICw&usg=AFQjCNEurxLz9LWB_PSfbxwuVmk-_zsd0g&bvm=bv.79184187,d.aWw

 

http://www.flightsim.com/vbfs/showthread.php?269153-Out-of-Memory-Problems&p=1796341#post1796341

 

 

...and a recent thread on tweaks for only the very adventurous to test in P3D:

 

http://forum.avsim.net/topic/453741-texturemaxload-27-instant-texture-loading/page-2

 

 

 

PS: When sharing "saved flight" scenarios with others to ensure that all possible FS "flight" parameters are identical on another person's installation of FS when attempting a comparison, be sure to include both the *.FLT (flight) and *.Wx (weather) files in an attached / linked *.ZIP file.

 

The *.FLT, *.Wx, (and *.SPB for missions) files used for "saved flight" scenarios are stored in ex:

 

C:\Users\[user account or profile name]\Documents\Flight Simulator X Files\

 

...and/or

 

C:\Users\[user account or profile name]\Documents\Prepar3d Files\

 

 

FYI: The importance of "saved flights" if doing comparison tests was recently discussed here:

 

http://forum.avsim.net/topic/453065-i-wonder-whysuch-great-global-130-performance-in-fsx-but-not-p3dx/page-2#entry3092761

 

 

 

NOTE: If other users do not have the same aircraft as was used in the original flight which was saved by the creator of that saved flight, that's OK ...as FS will simply substitute ex: the default C-172 aircraft for the other user(s).

 

 

BTW: The P3D SDK docs for "Flight Files" is here:

 

http://www.prepar3d.com/SDK/Mission%20Creation%20Kit/FlightFiles.htm

 

 

Happy tinkering ! :smile:

 

GaryGB

[n4gix 4,947]

Folks, you do realize that this parameter is pretty much useless for most machines these days.

 

"...on multi-core machines, there’s very little reason to tweak the fraction because it really only impacts performance of single core machines."

 

See Phil Taylor's Blog post: http://blogs.msdn.com/b/ptaylor/archive/2007/05/11/tweak-of-the-week.aspx

 

Also, read this carefully to understand precisely what this parameter does on single-core machines:

 

 

FIBER_FRAME_TIME_FRACTION determines the maximum amount of time per frame that we will run fiber jobs on the primary thread. We measure how long it took to simulate and render and then multiply that time by FIBER_FRAME_TIME_FRACTION to determine how long to run the fibers.

 

For example, if it took 10 milliseconds to simulate and render and FIBER_FRAME_TIME_FRACTION was set to the default value of 0.33, then we would allow the fibers on the primary thread to run for up to 10 * 0.33 = 3.3 milliseconds. For fraction values of 1.0 and 2.0, the time given to fibers would be 10 milliseconds and 20 milliseconds, respectively.

 

 

Now unless one is living in an alternate universe where the rules of math are different, then multiplication by zero still results in ZERO!

 

Therefore a value of 0.0 logically is simply killing the primary thread's fibers...

 

But, perhaps the most relevant quote from that article is this, from the mouth of Adam Szofran:

 

 

Please emphasize the pointlessness of tweaking this value on multi-core machines.

[simmerhead 885]

Then again, LM has encouraged tweaking this parameter several times for those who are strugeling with blurries. But as I see it, v2.4 is very nicely optimized. I have a failry outdated computer, and v2.4 runs good considering all the eye candy I throw at it. I do see blurries though in mountain areas with 10m mesh, and fairly often I get roads and railroades blurred.

[dukeav 46]

Folks, you do realize that this parameter is pretty much useless for most machines these days.

 

"...on multi-core machines, there’s very little reason to tweak the fraction because it really only impacts performance of single core machines."

 

See Phil Taylor's Blog post: http://blogs.msdn.com/b/ptaylor/archive/2007/05/11/tweak-of-the-week.aspx

 

Also, read this carefully to understand precisely what this parameter does on single-core machines:

 

 

 

Now unless one is living in an alternate universe where the rules of math are different, then multiplication by zero still results in ZERO!

 

Therefore a value of 0.0 logically is simply killing the primary thread's fibers...

 

But, perhaps the most relevant quote from that article is this, from the mouth of Adam Szofran:

And killing fiber' has what consqunces on P3D? I see none for me with locked frames.

 

You are quoting years old comments for FSX. With the results I have seen, I can safely say they no longer apply for P3D.

[GaryGB 10]

Hi Fr. Bill:

 

Thanks for pointing out that technicality; your insight as an "FSInsider" is always appreciated.

 

 

I believe is important to encourage learning options for optimizing one's FS experience in P3D. :p0304:

 

 

Futhermore, with changes LM has already made in the MSFS code base, it is possible that P3D may now behave differently than in FS9 or FSX when such FS tweaks are used, either directly because of the change in a tweak parameter value, or in-directly ...as a result of the 'interaction' that a tweaked parameter with other tweaked parameters.

 

 

Given the reported extensive scope of changes LM endeavors to implement while overhauling the MSFS code base and porting it to Dx-11 (...and DX-10 although that was less well-publicized), it is apparently going to be a "Work In Progress" (aka "WIP") for some time to come.

 

Whereas FSX was 'allowed' time only for SP1 and SP2 beyond being IMHO "shackled" to an incomplete preview showcase of the DX-10 'carrot' for a Windows Vista release, LM may be revising its code constantly over the coming years as it finds new ways to improve the FS code base (...rather than merely "putting lipstick on a pig").

 

So, IMO, only those with patience and a willingness to endure a 'WIP', and those willing to tweak for usability are likely to start and continue ...to use Prepar3d.

 

 

So, rather than seeing LM fail to achieve its 'mission' of getting greater numbers of the existing and new FS Community members to begin (...and continue !) using Prepar3d, IMHO, learning how to "tweak" in order to make the most of the 'currently coded' P3D is inevitable, if not "mission-critical" to success. 

 

I think this was at least part of the motivation behind the extensive posting of FSX tweaks by Phil Taylor of ACES at a time when the FS Community was clamoring wildly for better FSX stability and performance.

 

 

Otherwise, many prospective P3D users (and would-be P3D developers weary of having to be a "gamer" well-practiced in shooting at a moving SDK target) ...may simply feel too "WIP-P3D" to endure the ongoing process of "code-tweaking" that the LM team appears to be doing since the RTM roll out of what some might regard as a "not completely ready for prime time" product. :blush2:

 

 

Well, anyway, thanks again; and now I guess its time for another "test flight" !  :p0503:

 

GaryGB

[Guest]

100% with you Duke. There is no question about it that FFTF=0 with the Appropriate locked FPS constitutes a significant performance increase for a few here in this thread.

In my case I have always used FFTF=0.10 in FSX and Prepar3D and found it helped smoothness. When I saw a post with FFTF=0.01 I thought it was a typo but was informed that it was not. Tried a few different setting over a couple of days as posted in this thread and I am certain that FFTF=0.01 at unlimited FPS has made a big difference on my system and not only to performance, It is also more VAS friendly than whatever the default setting is or FFTF=0.10 and so is unlocked FPS more VAS friendly.

 

Here is my first somewhat proper flight with these setting.

Note the greedy addons, the very high setting, the overcast ASN weather and the FPS and the astonishing absence the Ping of death and the OOM that would have been inevitable on the approuch everytime I tried to do this before. This time I make it all the way to the gate. The only thing that I forgot to do was to reset the flight which usually results in an OOM if it was imminent anyway.

If I had gone to the scenario screen and loaded a different flight or reset the same flight without getting an OOM that would have been very telling IMO:

 

[BimmerCop 332]

Folks, you do realize that this parameter is pretty much useless for most machines these days.

 

"...on multi-core machines, there’s very little reason to tweak the fraction because it really only impacts performance of single core machines."

 

See Phil Taylor's Blog post: http://blogs.msdn.com/b/ptaylor/archive/2007/05/11/tweak-of-the-week.aspx

 

Also, read this carefully to understand precisely what this parameter does on single-core machines:

 

 

 

Now unless one is living in an alternate universe where the rules of math are different, then multiplication by zero still results in ZERO!

 

Therefore a value of 0.0 logically is simply killing the primary thread's fibers...

 

But, perhaps the most relevant quote from that article is this, from the mouth of Adam Szofran:

 

I do realize that Phil Taylor and Co, were talking about FSX and not P3D v2.4... LM has done extensive changes to the core ESP engine and have even adviced to tweak the FFTF in their tuning guide. If they say this setting can be used to tweak P3D, they must have a good reason for it. I doubt they would suggest the use of a tweak if it only applied to single core CPUs which are hardly used nowadays.

 

In MY case, FFTF = 0.001 and FPS locked at 40 has given me a tremendous boost in performance. And no, it's not snake oil or placebo. I have not changed anything on my setup other than adding FFTF = 0.001.

 

Perhaps the fact that I don't use autogen (MSE user here) gives me an edge and the ability to use such FFTF tweak. Who knows. But I can say that while I used to average low 20s taxiing around Taxi2Gate's MCO scenery with FFTF = 0.33, I now am averaging low 30s, even with ASN and heavy clouds.

 

As a matter of fact, I was able to slide many options full right, enable cloud shadows, HDR, etc, thanks to this FFTF tweak. And IQ has not suffered one bit. My terrain is sharp all the way out to as far as I can see. No blurries, no texture swapping.

 

So to say this tweak is pretty much useless on P3D v2.x machines is far from the truth.

[FlyIce 354]

Dave, very nice video and I'm astonished that you slided pretty much all graphic settings to the max and still managed to get staying above 24fps most of the time with PNW over a mega city. That's unbelievable if not unthinkable. I almost gave up on Orbx PNW but after seeing your video I want to revisit it with FFTF=0.01.

 

I checked your profile but didn't see your system info. Would you mind to share what kind of beast system you're running? I just received my GTX980 today, and will try your settings tonight see what fps I can get. I am already running 780ti so don't know how much increase I may see from 980..

[DylanM 1,069]

You are quoting years old comments for FSX. With the results I have seen, I can safely say they no longer apply for P3D.

 

I can't agree more on how much this is not a placebo, and how things have evolved since that blog was posted...

[Sesquashtoo 2,846]

That's the whole point. It is possible to make 0.001 work without  getting blurries and losing autogen, but then the advantage of improved frame rates is lost by using the limiter. I guess that tells us that these items are interlocked and that's probably how the Fiber Frame Utility works. It adjusts FFTF on the fly to get the target frame rate that your specify.

This works....  with the setting of 0.01 and a locked in-sim setting of 33 FPS.  I have been running around Palm Springs all day at 30-33 FPS, no blurs, and this is better than what I was getting with unlimited. With unlimited, it banged around between 22-27-33 FPS with an external limiter.  I have now turned that off, and am using the 0.01 setting, and limited internally to 33 FPS.  It is working great!  Good tweak,  and the only one I have set into place for P3D.

[Guest]

 

 

I checked your profile but didn't see your system info.

 

Well, its there now. Although I spared no expense building it and apart from getting very luck with a CPU that can do 5ghz under the right conditions, its no that different from the setup of many who often ask on my Youtube channel why don't I get that performance from my virtually identical setup? I always answer that I followed NickN's Haswell and FSX bible guides Meticulously, plus my cooling is somewhat novel.

 

 

 

with the setting of 0.01 and a locked in-sim setting of 33 FPS.

 

Do you mean 0.001?

I found that 0.01 was better with unlocked FPS

 

0.001 absolutely requires that you lock the FPS!

[FlyIce 354]

Thanks! I think my system overall is comparable to yours (2500K @4.9GHz) + 780ti + SSD etc. I'll copy exactly your settings and see what kind of performance I can get. I'll be amazed if I can maintain above 24fps with many sliders to the right most and so many boxes checked.

[Sesquashtoo 2,846]

Well, its there now. Although I spared no expense building it and apart from getting very luck with a CPU that can do 5ghz under the right conditions, its no that different from the setup of many who often ask on my Youtube channel why don't I get that performance from my virtually identical setup? I always answer that I followed NickN's Haswell and FSX bible guides Meticulously, plus my cooling is somewhat novel.

 

 

 

 

Do you mean 0.001?

I found that 0.01 was better with unlocked FPS

 

0.001 absolutely requires that you lock the FPS!

Actually, referring to 0.01 with my locked 33 FPS.  I'm getting top FPS hits at this setting.  With unlimited, it went through a range of 22-33 FPS.  With it locked at 33, it now never goes below 27/low 30 median/-33 at top limit.   It's a keeper. I wonder if this would also play out with FSX? 

[newtie 533]

I wonder what LM has to say about all this.