i7 here, i3770k @ 4.6ghz, without AM, using FFTF=0.01 twak only, and locket internally at 60 - P3Dv3 smooth as silk, 50-60fps at max settings, around 30-40 with the NGX. GTX 960 2gb is the GPU. So, as far as I can see, P3Dv3 is another FSX, behaving weird on different systems. Correct me if I'm wrong.

HT on or off?

 

Remarkable thread, thanks so much, Daniel.

 

Kind regards, Michael

Good stuff Daniel,  must take a closer look at this over the weekend 

You're welcome, everyone.

 

I'm quite impressed by the Fiber Frame Time Fraction = 0.01 tweak. After reading it shouldn't make a difference for multi-core machines, I had disregarded it. But now I see the results firsthand. It does certainly make a quantifiable difference. Here I can now push the internal frame lock to 35 FPS while keeping LOD_Radius at Max.

 

AF = 85, 35 FPS Upper Limit, HT on, Overclock to 4.7 GHz, FFTF = 0.01

 

 

It's still remarkably smooth. 99.9% of all frames are rendered > 20 FPS, 99% > 25 FPS, and the average is right at the upper frame limit. The FFTF tweak does actually work for multi-cores.

HT on or off?

 

Remarkable thread, thanks so much, Daniel.

 

Kind regards, Michael

 

HT on.

Thank you Daniel for taking the time to prepare such a detailed analysis and make available to us.

 

I had made the decision when I moved to P3D V3 that I was going to resist any changes from the standard. However as you say the sim was unusable from a fluidity point of view.

 

I really like what LM have done with V3, it is an awesome looking sim. Most people are reporting better performance than V2.5. Unfortunately I do not fall into that camp. I am struggling to achieve the same level of fluidity. The FTTF=0.01 has provided the most benefit to me. It has returned valuable FPS and CPU overhead to me to increase fluidity (I lock internally at 33FPS).

 

Anyway off to play with the AF some more to see what I can improve on.

 

Thank you

Michael M

You're welcome, everyone.

 

I'm quite impressed by the Fiber Frame Time Fraction = 0.01 tweak. After reading it shouldn't make a difference for multi-core machines, I had disregarded it. But now I see the results firsthand. It does certainly make a quantifiable difference. Here I can now push the internal frame lock to 35 FPS while keeping LOD_Radius at Max.

 

AF = 85, 35 FPS Upper Limit, HT on, Overclock to 4.7 GHz, FFTF = 0.01

 

It's still remarkably smooth. 99.9% of all frames are rendered > 20 FPS, 99% > 25 FPS, and the average is right at the upper frame limit. The FFTF tweak does actually work for multi-cores.

 

 

Fiber Frame Time Fraction = 0.01 is supposed to be an invalid value and the sim consequently defaults to 0.1.

Would be interesting to compare 0.01 with 0.1 here.

 

gb.

Here's a comparison with the same setup in FSX:SE.

 

AF = 85, 30 FPS Upper Limit, HT on, Overclock to 4.7 GHz, LOD_Radius = 4.5, FFTF = 0.01 | FSX Steam Edition

 

 

Pretty good, but there is more frame time instability. Prepar3D wins here.

Fiber Frame Time Fraction = 0.01 is supposed to be an invalid value and the sim consequently defaults to 0.1.

Would be interesting to compare 0.01 with 0.1 here.

 

gb.

 

I have a feeling the differences would be indiscernable. But I'll give it a try.

Great thread Daniel!  Love FreeMeshX and use all the time, great experience by the way.

 

FFTF=0.01 has been my single best tweak in both P3D v2 and v3.  Bar none nothing else helped more.  Basically it always gave me the advantage the increased performance of "unlimited" while the smoothness of a "limited" framerate.  Had a long discussion about this previously.

 

I think SSD's help here, as others report textures loading late when using HDD's.

Great job!

 

Without FPS lock, FFTF=0.1 can cause severe loss of autogen. Without FFTF=0.1, FPS lock will cause severe loss of FPS. Now you have the quantitative data to prove FFTF=0.1 + FPS lock indeed is a perfect combination.

Great thread Daniel!  Love FreeMeshX and use all the time, great experience by the way.

 

FFTF=0.01 has been my single best tweak in both P3D v2 and v3.  Bar none nothing else helped more.  Basically it always gave me the advantage the increased performance of "unlimited" while the smoothness of a "limited" framerate.  Had a long discussion about this previously.

 

Awesome! Glad to hear!

 

I wanted to test out this theory to see if FFTF = 0.01 gives any stability to a frames unlimited state. So without further ado,

 

AF = 85, Frames Unlimited, HT on, Overclock to 4.7 GHz, FFTF = 0.01

 

 

Amazingly, the micro-stutters associated with unlocked frames almost completely disappear with the FFTF tweak. However, we are still subjected to a few long frames, but only very few. In comparison to the locked frame rate of 35 FPS experiment, the average frame rate does increase from 34 to 37 FPS, the minimum 1% increases from 25 to 26 FPS, but the minimum 0.1% drop to 17 FPS from 21 FPS, indicating unlocked frames does still add some stuttering. But not nearly as severe as before. There is merit to unlocked frames if the FFTF tweak is applied.

 

I didn't have a chance to observe autogen issues. A five minute test probably is not enough time to ascertain whether autogen rendering is affected with the FFTF tweak.

 Setting process priority to "high" for prepar3d.exe makes quite a difference on fluidity on my system. 

 

I did this for years with FSX and used Prio to do it. Never tried it with P3D.

O.K, ...stupid question time...

 

where do you set the affinity mask, prepar3d.cfg, Taskmgr, or some other place?

 

what are acceptable comment chars in .cfg files? //,;,;;,*/...?

 

what is the right syntax, for setting Affinity Mask, AF=?

 

Thanks,

 

Chas

O.K, ...stupid question time...

 

where do you set the affinity mask, prepar3d.cfg, Taskmgr, or some other place?

 

what are acceptable comment chars in .cfg files? //,;,;;,*/...?

 

what is the right syntax, for setting Affinity Mask, AF=?

 

Thanks,

 

Chas

In the prepar3d.cfg

 

add:

 

[JobScheduler]

Affinitymask=XX

If i use frames unlocked, it gives butter smooth performance in near all conditions, however, my GPU usage skyrockets....especially in clouds it goes to 99% !!

I would like to see some test with a 6-core processor like a 5820 or 5930, 5960....