What's the matter with you Dirk? It's all been mentioned before, you're not keeping up.

Low fps if you want FFTF controlled externally, and that is more suited to low core counts because higher core count used for the sim reads in the data fast enough anyway. OK?

...before you ask, you're going to want to know what a low core count is and what a high core count is?

A low count is that which struggles with the background task, a high core count is that which does not. 

10 minutes ago, SteveW said:

What's the matter with you Dirk? It's all been mentioned before, you're not keeping up.

Low fps if you want FFTF controlled externally, and that is more suited to low core counts because higher core count used for the sim reads in the data fast enough anyway. OK?

Got it, "best" was in the sense of suited, thanks

Cheers

Let's run over it quickly:

I used a descent route into EHAM from a low complexity at 4000ft to a high complexity at 1200ft over four minutes.

The same test with a setting of 0.05 in the cfg looks very similar but shows less loading speed at the start. So we know FFTFd has set around 0.33 there at the start. At the end we have the same look as setting 0.05 in the cfg, so we know FFTFd has moved down to near 0.05 by the end.

What I noted was that there is no apparent deterioration to the flow with FFTFd in operation.

The greater loading 0.33 at the higher altitude serves a use as we gain velocity and encounter more tiles.

I simulated a four core running P3D v4.2, and set aside the running apps on two cores. I used HT enabled to reduce switching across the CPU. I think it was a good test to show how it works but it does not show if it can improve the performance in certain areas in your sim.

 

...Some of you guys are troubled by the way I've set up for the test and commented on how it doesn't suit their way of flying to set up like that. Again in the Core Affinity section this is a setup to aid the fair result of the test and in no way represents the best way to set up your simulator. Unfortunately I don't have time to bring everyone up to speed on scientific techniques.

The test was designed to *surface* a trend that reveals the operation of a thing, it is not to find the best way to set up the simulator.

 

Edited April 3, 20188 yr by SteveW

1 hour ago, SteveW said:

A low count is that which struggles with the background task, a high core count is that which does not. 

My 8LPs on 6700K @4.6GHz AM 243 are enough @unlimited FPS/25Hz with my P3Dv4.2 settings in the most conjested areas like AS Heathrow with ORBX England, so FFTF Dynamic wouldn't help me here, but I'm still curious if I just don't know exactly what the blurries look like from the crz flight level, where this app could be helpful if I had those, of which I'm not sure ))

Edited April 3, 20188 yr by Dirk98

Understanding the possible advantage of the dynamic FFTF technique - think getting lower and slower, exchanging the power to render view distance with the power to render nearby detail.

3 hours ago, Dirk98 said:

Got it, "best" was in the sense of suited, thanks

Cheers

I don't mind leg-pulling for fun. : )

If you are not sure what I mean in a short post scroll back, I'm usually adding to what I've already said rather than repeat things for clarity - especially when I'm busy. Good sport.

41 minutes ago, SteveW said:

Understanding the possible advantage of the dynamic FFTF technique - think getting lower and slower, exchanging the power to render view distance with the power to render nearby detail.

That's exactly what I've gleaned from the discussion.

See, just with my config described above, and for those like me, who are many, it is rather vice-versa task than you describe, as I know that @Unlimited with my current hardware and software settings I get a very satisfactory rendition in 'lower and slower' highly congested areas. So I need to look more into what I don't like at high altitudes, if anything yet. I hope you understand what I mean.

Thanks,

Edited April 3, 20188 yr by Dirk98

Well, it also recovers the higher FFTF value as you get higher where you get faster so the converse is also true.

Unfortunately it is not easily predictable the advantage any one system can obtain from it, you would have to try it.

Ok  might be stupid, but i dont get all the differences between the FSPS products, who can explain the differences in what they do?  and not it is automatic or not

1) fiber accelerato2

2)2018 Live

3)FFTF

20 minutes ago, l_schuiling said:

Ok  might be stupid, but i dont get all the differences between the FSPS products, who can explain the differences in what they do?  and not it is automatic or not

1) fiber accelerato2

2)2018 Live

3)FFTF

Fiber Accelerator —> pre P3Dv4 

Booster Live —> you manually change the FFTF value ( possible when flying )

FFTF Dynamic —> Booster Live but automaticallly and AGL / Framerare mode 

FFTF Dynamic is what you want ( better than FA )

so FFTF is doing exactly what 2018 is doing and in essence the same as fiber accel?  so it is not only changeing the FFTF value but a bit more on the background?

 

regards

Leon

Another aspect of the test which I have not discussed is that during the entirety of the test, the test harness moves 90 objects a second around to stress the SimConnect subsystem. So you can forget worrying about your sode installation.

28 minutes ago, l_schuiling said:

so FFTF is doing exactly what 2018 is doing and in essence the same as fiber accel?  so it is not only changeing the FFTF value but a bit more on the background?

 

regards

Leon

FFTF Dyn controls the FFTF value , just like Booster 2018.

Fiber Acc has a slightly difference approach. 

Imho FFTF Dyn is the best...