FFPS:FFTF question

[Guest]

It took me years and years before I suddenly got how VORs work. Somehow it never clicked with me. But suddenly I got it and I thought: this is so simple, how could I not get it sooner?!? Well, I have the same feeling with AM and cores... I read about it but I never get it. It's a mystery to me. While I have the idea it must be really simple once you get it. I doubt if I will ever get it though. I've read 'thousands' of posts about AM and cores, including the ones in this topic, and I still don't get it. Those weird random looking numbers, those 000111010110 or whatever, it still boggles my mind...

[Victoroos 205]

40 minutes ago, J van E said:

It took me years and years before I suddenly got how VORs work. Somehow it never clicked with me. But suddenly I got it and I thought: this is so simple, how could I not get it sooner?!? Well, I have the same feeling with AM and cores... I read about it but I never get it. It's a mystery to me. While I have the idea it must be really simple once you get it. I doubt if I will ever get it though. I've read 'thousands' of posts about AM and cores, including the ones in this topic, and I still don't get it. Those weird random looking numbers, those 000111010110 or whatever, it still boggles my mind...

Haha, yepp. And then they start talking about binaries and spreading the load etc.. I always think of bread, a knife and a lot of peanut butter #pindakaas . 

[HUSSAR 262]

5 hours ago, skysurfer said:

Update. After running CCCleaner and reinstalling Client I was managed to start P3D. I started P3D several times and P3D stared OK. After that I installed FFTF back again P3D fails to start again. After uninstalling FFTF P3D failed too start. This is definitely FFTF causing P3D to fail to start!!! I am gonna call for refund. 

 

Hi Edward,

Are you installing using full admin rights and not installing from a separate drive?  Did you purchase the version for Prepar3Dv3 by mistake?  Highly unusual for this utility not to work on your system.  

[Achilles Philippopoulos 115]

8 hours ago, joepoway said:

Not to beat a dead horse as they say but from what I think you are saying is this:

If we set Unlimited frames in P3D no matter what entry we put in the config file for FFTF it is ignored and does NOTHING and the sim runs as though it were 0.01?

But if we run Locked frames with no entry the default is 0.33 but IF we add an entry into the config file the new value will be used (obviously not dynamically).

So therefore FFTF entries into the config file are only applicable when using Locked Frames (something I never knew)

HOWEVER through some "tricks" FFTF altered values do work in Unlimited mode (and obviously locked frames mode) with this Dynamic app.

So your bottom line is if you want "modified" FFTF values when running Unlimited frames it's only possible with this app and we are accomplishing nothing by changing that value in a config file mod IF we run Unlimited frames, Correct?

Joe 

3

Hi Joe,

>>If we set Unlimited frames in P3D no matter what entry we put in the config file for FFTF it is ignored and does NOTHING and the sim runs as though it were 0.01?<<

This is what I see from the LM code. The main part that contains FFTF value is not being called, or it is being executed so fast that cannot be caught. But I also suspect that another thread is running in this case (did not find it yet) because of some other findings related to frames settings.

>>So therefore FFTF entries into the config file are only applicable when using Locked Frames (something I never knew)<<

Yeap.

>>HOWEVER through some "tricks" FFTF altered values do work in Unlimited mode (and obviously locked frames mode) with this Dynamic app.<<

Yeap. That's why it was in beta testing for over 2 months.

>>So your bottom line is if you want "modified" FFTF values when running Unlimited frames it's only possible with this app and we are accomplishing nothing by changing that value in a config file mod IF we run Unlimited frames, Correct?<<

Yeap. The part code looks to be ignored, so we have to force it not to be ignored somehow :-)

 

 

 

   

 

[skysurfer 76]

It says for P3D 4.

I spent many hours to make P3D to work and I still cannot start it. I was able recover it but after reinstalling FFTF it stopped working again. I am completely puzzled. I tried to reinstall client for the second time but P3D still won’t start. I am waiting for P3D to respond 

[Achilles Philippopoulos 115]

6 hours ago, SteveW said:

Remember I said that that the small climb in fps with high FFTF values is due to less 'linkage' time the main thread is stopped a lot less often.

The time taken to load the sim shows clearly the differences that we get with changing the FFTF value set in the Prepar3D.cfg

Here are the graphs captured of those simple changes to FFTF on a system guaranteed to reproduce the same scene each rerun.

 

Clearly setting FFTF to 0.9 loads the scene fastest and we can see a small rise in fps, 0.1 takes ages. Sure enough looks like the 0.33 value is the same as no value. This is in Unlimited, no Vsync, No TB. Misty rainy view cross cockpit to tower, lots of clouds thunder, light buildings and autogen.

 

Hi Stewe,

Does the above results being produced in playback mode?

[HUSSAR 262]

2 minutes ago, skysurfer said:

It says for P3D 4.

I spent many hours to make P3D to work and I still cannot start it. I was able recover it but after reinstalling FFTF it stopped working again. I am completely puzzled. I tried to reinstall client for the second time but P3D still won’t start. I am waiting for P3D to respond 

Hopefully Mr. Phiiippopoulos will reply and help you out.

[Nemo 1,305]

4 hours ago, J van E said:

It took me years and years before I suddenly got how VORs work. Somehow it never clicked with me. But suddenly I got it and I thought: this is so simple, how could I not get it sooner?!? Well, I have the same feeling with AM and cores... I read about it but I never get it. It's a mystery to me. While I have the idea it must be really simple once you get it. I doubt if I will ever get it though. I've read 'thousands' of posts about AM and cores, including the ones in this topic, and I still don't get it. Those weird random looking numbers, those 000111010110 or whatever, it still boggles my mind...

Whenever there is a topic about a specific p3d tweak,no matter which, there is always a good chance that AM discussion comes in. Although, there might be good reasons for this AM discussions and suggestions, in my case, I could never see a significant benefit in using AM tweaks.

[Ted Striker 338]

5 hours ago, J van E said:

It took me years and years before I suddenly got how VORs work. Somehow it never clicked with me. But suddenly I got it and I thought: this is so simple, how could I not get it sooner?!? Well, I have the same feeling with AM and cores... I read about it but I never get it. It's a mystery to me. While I have the idea it must be really simple once you get it. I doubt if I will ever get it though. I've read 'thousands' of posts about AM and cores, including the ones in this topic, and I still don't get it. Those weird random looking numbers, those 000111010110 or whatever, it still boggles my mind...

I agree with this 100% except with me it was trying to understand derivatives in first year calculus.

I see a lot of rambling and theories about AM and FFTF in the forums but I have yet to see a clear explanation of how the basics of the sim works with regards to these settings that I can understand. Until the knowledge can pass through my thick skull and the light bulb comes on in my head I am staying with my 4 core HT off, no AM which has been working just fine since I built my system per Nick Needham's guide back in 2013.

For example. It has been stated numerous times in this thread that FFTF has no effect with unlimited frames. The question has been asked numerous times as to whether that applies to unlimited with vsync but that has not been answered. Then Steve posts his charts showing the different results of various FFTF values and he is running unlimited frames which is not supposed to work. No wonder we are confused.

Ted   

[joepoway 241]

9 hours ago, SteveW said:

Clearly setting FFTF to 0.9 loads the scene fastest and we can see a small rise in fps, 0.1 takes ages. Sure enough looks like the 0.33 value is the same as no value. This is in Unlimited, no Vsync, No TB. Misty rainy view cross cockpit to tower, lots of clouds thunder, light buildings and autogen.

 

Steve

Thanks for staying engaged, always welcome.

It would seem you are saying that FFTF changes when using unlimited mode do make a difference based on load times mentioned above since you say “This is in Unlimited” meaning turned ON?

So if those changes are recognized during load times wouldn’t they also effect the percent dedicated to terrain load times during flight or is that completely different?

There seemed to be confirmation of my summary regarding what doesn’t happen if Umlimied is deployed but now I’m not sure based on your statement above.

At this point it’s a matter of just trying to understand for my own sake when FFTF is deployed and what other factors or settings make it a mute point. I believe I know what it does but now I’m a bit confused when it’s allowed to do its thing which makes performance testing by using it difficult to say the least.

Thanks in advance

Joe

[onebob 125]

1 hour ago, Ted Striker said:

but I have yet to see a clear explanation of how the basics of the sim works

Here goes, for FSX sp2 the explanation I came across for a 4 core cpu:-

For a frame rate  say 30

1000/30= 33.33m's/frame

for a FFR value of 0.33= 33.33x.33=11m's  ( why .33 default, is was sugested after testing a value of 10m's was the best result)

in the 4 core cpu

Core 0 is responsible for Fibers
Core 1 is responsible for Main Sheduler

core 2 & 3 are responsible for Texture Management and Object ( Autorgen) batching.

therefore :----

input into Fibers                 ( at 33.33m's)  ...core 0  ############ .Fibers must co-operate with  -------->

Main Sheduler .....( at 11m's I think ?)..........core 1  ############  Main Sheduler Renders --------------->

-------------------------------------->                 .....core 2  ############
                                                                .....core 3  ############

another value that comes (some were) into the equation is:-

SWAP_WAIT_TIMEOUT=30
This variable is the number of frames that the terrain engine will wait for terrain textures to be loaded into video memory before forcing new tiles of terrain triangles to be rendered.
in dx10 it was suggested values 10 or below

Now we are into 4 cores and more and also with HT and with LM moveing more on to the additional cores, The AM will move the above work on to cores other than core 0

bob

Edited March 25, 2018 by onebob

[Guest]

1 hour ago, onebob said:

Here goes, for FSX sp2 the explanation I came across for a 4 core cpu:-

You lost me when you said 'FSX sp2'. Well, I did read the rest but it didn't make things easier. On the contrary. Thanks for the effort though.

It's not that I actually want to tweak btw. It's just that I read so much about AM and HT and cores and LP's but never really understand it all. Yes, I do understand the concept but that's it: I don't get mathematics and numbers and what to actually do. But it's not a problem, not at all: I am perfectly happy with performance here (i7 4790K@4.5, no HT, no AM, no tweak at all). Obviously it does help that the only plane I fly, the Majestic Q400, is a real wonder when it comes to 'performance of study level aircraft'.

[Ted Striker 338]

2 hours ago, onebob said:

Here goes, for FSX sp2 the explanation I came across for a 4 core cpu:-

Bob,

I appreciate you taking the time to try and explain but you lost me with that.

What is the 1000 in "1000/33" ?  What is a m's?

I understand that with .33 FFR that .33 of each cycle is spent on loading or rendering something. Whether it is mesh, terrain tiles, scenery objects, autogen, etc. I am not sure. Then, when does it work and when does it not? Very conflicting opinions in this thread about that.

"Main Sheduler .....( at 11m's I think ?)..........core 1  ############  Main Sheduler Renders --------------->"   Huh? Is this computer code?

SWAP_WAIT_TIMEOUT=30  Just the name of that one sounds like it is intended to confuse! I know this one has been around for awhile, but what does it do and how should it be used with FFTF? How should they both be tested and monitored?  A value of 10 was suggested. Under what circumstances?

Affinity Mask is the most confusing. Again, many conflicting statements and theories on how this works. Perhaps you just need to be a computer engineer to understand it. It does not seem to be as simple as just assigning the sim and each addon to dedicated processors. Apparently a lot of trial and error testing has to be done. I have not seen a systematic approach to how this testing should be done and what tools should be used to monitor it.

I realize that this is an open forum and everyone is volunteering their time so I certainly do not expect anyone to provide the details that I am describing. I am just saying that to me a lot of this sounds like guesswork theories, no different than we've had for the last 3 decades of flight simulator tweaking.   I hope no one is offended by my lack of ability to fully understand this AM and FFTF stuff. I am not complaining, my sim works fine as it is without it.

It would really be great if Lockheed Martin would host a few lecture sessions at one of the Flight Sim expos this summer to discuss topics like these and other inner workings of P3D.

I must admit that I have not read the SDK, so maybe I need to start there! Perhaps its as simple as reading the manual!

Ted

 

Edit: I just wanted to return and apologize for derailing this thread. Gerard has spent a lot of time testing the FFTF tool, is very satisfied with it, and started this thread to share his enthusiasm. The last thing I want to do is derail this to yet another Affinity Mask discussion. Let's get back to the FFTF tool.

Ted

Edited March 25, 2018 by Ted Striker
Thread derail apology

[SteveW 1,792]

9 hours ago, Achilles Philippopoulos said:

Hi Stewe,

Does the above results being produced in playback mode?

Yes, straight run of the sim with FFTF value set in the cfg. Three goes with each setting. Sim set to Unlimited on the slider, VSync=Off, TB=Off.

With FFTF=0.1 the scene around the player tales longer to fill in. The point at which the sim starts for each is roughly the same within a second or so. The time to take to collect the scenery around the player extending out from the player is what we can see taking less time with 0.9

It's done scientifically with the use of a test harness, so no mistakes.

 

The problems I see regarding understanding these concepts is that far too much rubbish is spoken about AMs HT FFTF and so on from those that are merely guessing.

[Achilles Philippopoulos 115]

20 minutes ago, SteveW said:

Yes, straight run of the sim with FFTF value set in the cfg. Three goes with each setting. Sim set to Unlimited on the slider, VSync=Off, TB=Off.

With FFTF=0.1 the scene around the player tales longer to fill in. The point at which the sim starts for each is roughly the same within a second or so. The time to take to collect the scenery around the player extending out from the player is what we can see taking less time with 0.9

It's done scientifically with the use of a test harness, so no mistakes.

 

The problems I see regarding understanding these concepts is that far too much rubbish is spoken about AMs HT FFTF and so on from those that are merely guessing.

Hi Steve,

In playback mode, FS code for FFTF works at unlimited frames but it works from another FS code routine.

At playback mode, you need high FFTF value because otherwise you see thinks like it was at FS2002 :-)

I also implemented it, to support simmers for video production. 

Edited March 25, 2018 by Achilles Philippopoulos