FSX.CFG "Tweak Test" performed today

[Guest Raul_Hernandez]

Hi all,

 

I offer this as nothing more than anecdote or narrative of some tests (/playing around) I have been doing this morning concerning tweaks to the FSX.CFG. :smile:

 

The results were interesting, and not at all what I expected.

 

The Test Bed

 

- I am using a stock i7 2600k 3.8Ghz, with 8GB DDR3 RAM, a ATI HD 5770 1GB, and Win 7 64 bit

 

- All FSX settings remained identical for all 4 tests. I won't bore them with you, but they propably average settings, eg.. Texture Complexity: Very Dense, Autogen: Dense, Airline AI: 20%, Road Traffic: 15%, etc, etc

 

- ATI CCC set at AA of 4x and AF 4x and other 'average' type settings.

 

- PMDG 737 NGX with all 'performance manager' settings on "High Resolution"

 

- Loaded on the Runway 16 threshold of FlyTampa Vienna v2.

 

- I am an OPUS user but for this test, used FSX Standard Weather theme "Fair Weather"

 

 

Scenario 1 :

 

Totally new, fresh, standard FSX.CFG with no tweaks whatsoever

 

Scenario 2 :

 

Totally new, fresh, standard FSX.CFG with [HIGHMEMFIX] tweak ONLY.

 

Scenario 3 :

 

Totally new, fresh, standard FSX.CFG with all recommendations from Word Not Allowed's guide applied (except GPU recommendations, due to being ATI)

 

Scenario 4 :

 

Totally new, fresh, standard FSX.CFG with all tweaks by Bojote's automated tweaks site applied.

 

 

Ok, for all scenarios and all tests, I ran with frame rates locked at 30 FPS, using the standard internal FSX frame limiter. This slightly goes against the Bojote advice which is to leave frames unlocked in FSX and use an external limiter, but in all other areas, all recommendations were followed.

 

I added the [Average Frame Rate] tweak to the cfg, in order to compare average frame rates. I left each scenario to sit on the runway for 60 seconds after loading, to allow for any 'stabilization' time, of frame rates.

 

OK, HERE'S THE RESULTS :-

 

SCENARIO 1 : Average Frame Rates = 23 (Totally Stock)

 

SCENARIO 2 : Average Frame Rates = 28 (High Mem Fix only)

 

SCENARIO 3 : Average Frame Rates = 25 (Word Not Allowed's Guide, where applicable.. eg ATI)

 

SCENARIO 4 : Average Frame Rates = 26 (Bojote's Automated Tweak Site)

 

SUMMARY

 

In this test, I was extremely surprised to find that a totally 'stock' FSX.CFG with nothing at all changed or added, other than the [HIGHMEMFIX] resulted in the highest frame rates. It is widely publicized that Word Not Allowed's guide makes references only to nVidia settings, so again it doesn't fit so comfortably within this test, however there are plenty of ATI users out there and I am sure many will be interested in all tweak approaches.

 

What I also noticed is that Scenario 2 was the smoothest and most fluid experience. Very small micro stutters existed in both Scenarios 3 & 4.

 

What surprised me most is that I thought the HIGH MEM FIX tweak was just to prevent OOM errors, and that it would not impact momentary performance at all. But with a 5 FPS (around 20%) performance increase in frames with the tweak applied, it was clear to me that this tweak does positively impact performance.

 

I am not offering any conclusions here other than to encourage everyone to do a similar test yourself - you, like me, may not get the results you were expecting to get!

 

I am now permanently using a completely 'untweaked' FSX.CFG except for the High Memory Fix. :wink:

[wanabflyer 53]

One Question: In all four scenarios, was your mouse pointer in the same location on your screen?

 

Regards,

[simmerhead 885]

Welcome to the club. HIGHMEMFIX is my only tweak also. It doesn't just reduce the risk of OOM errors, but enables FSX to use more graphics memory (>512 MB) when available.

 

Nothing wrong with tweaking, but don't do it if you don't know what the heck you are doing...

[Guest Raul_Hernandez]

One Question: In all four scenarios, was your mouse pointer in the same location on your screen?

 

Regards,

 

I know what you're getting at, as I had a 'mouse pointer' issue on my last rig, running a 32 bit OS, but in my current system, the position of the mouse pointer makes no impact whatsoever on performance. In any case, the answer is yes, broadly speaking each 'scenario' had the same periods with the mouse on screen, then off etc.

[wanabflyer 53]

I know what you're getting at, as I had a 'mouse pointer' issue on my last rig, running a 32 bit OS, but in my current system, the position of the mouse pointer makes no impact whatsoever on performance. In any case, the answer is yes, broadly speaking each 'scenario' had the same periods with the mouse on screen, then off etc.

 

Now there is something interesting. I have a system, not a heck of a lot different from yours, except i am OC'd at 4.4 and i have a Nvidia card and i still have an impact.

 

Regards,

[simmerhead 885]

 

 

Now there is something interesting. I have a system, not a heck of a lot different from yours, except i am OC'd at 4.4 and i have a Nvidia card and i still have an impact.

 

Regards,

 

Depends on aircraft used, doesn't it?

[Guest Raul_Hernandez]

Now there is something interesting. I have a system, not a heck of a lot different from yours, except i am OC'd at 4.4 and i have a Nvidia card and i still have an impact.

 

Regards,

Depends on aircraft used, doesn't it?

 

I think it does certainly depend on the aircraft used.... (iFly 737 for FSX was one of the only products I ever got a refund on from Flight 1, due to frame rates dropping from 30 to 8, when the mouse was in view)....

 

But also the last rig that I had a few 'mouse cursor in view' issues with, was using an old 8800 GTX card.... since getting my current rig with an ATI 5770 1GB I have never seen any mouse/FPS issues, on any aircraft.

 

Perhaps it is one area where nVidia has the issue rather than ATI ... (makes a nice change! :lol:)

[Gazzareth 808]

Just as a thought, was there any visual difference on the screen - blurries or sharpness issues either way? I am sure some of the tweak change that, which in itself would reduce the frame rate....

 

G

[MarkSC 122]

What about the AffinityMask tweak for quad core? I thought that was a must-have too? Basically, I think that today's CPUs and GPUs have finally about caught up to FSX.

[Guest Raul_Hernandez]

Just as a thought, was there any visual difference on the screen - blurries or sharpness issues either way? I am sure some of the tweak change that, which in itself would reduce the frame rate....

 

G

 

No discernible visual (blurry) differences between any of the scenarios. But Scenerios 3 & 4 had just a slight stutter when panning with TrackIR, that was not evident with Scenarios 1 & 2 (even considering Scenario 1 having the lowest frames).

 

 

What about the AffinityMask tweak for quad core? I thought that was a must-have too? Basically, I think that today's CPUs and GPUs have finally about caught up to FSX.

 

To be honest this was one of the main things I was looking at in doing this 'test', as an i7 user I wanted to see if there would be a big difference;

 

So Scenarios 1, and 2 (the highest frames) have no Affinity setting at all.

 

I've just flown my first full flight since adopting the "Scenario 2" CFG file; using the RealAir Turbine Duke, flying in FTX England, from EGNT to EGNJ. I done this same flight yesterday with a Bojote CFG in place ....... honestly I cannot believe how smoother and better frames it was today, with the "Scenario 2" CFG, and it was even raining (OPUS Weather) today, which it was not yesterday. Rain used to drive down frames with my previous CFG but with this new "au naturale" CFG, it doesn't seem to impact.

 

As always YMMV - I think a lot of what I am seeing is caused by the fact that all of the main tweak guides focus on nVidia, so perhaps they are actually counter-productive when used with ATI. With ATI, it seems that for me at least, just the [HIGHMEMFIX] is all that should be added to the CFG for best performance.

[simmerhead 885]

What about the AffinityMask tweak for quad core? I thought that was a must-have too?

 

Must have? Can't think of any reason for it, unless you run some very intensive programs simultaniously with FSX. If you run Windows 7 it will make FSX use all cores just fine. No two systems are alike though, so as with all tweaks they might or might not work on yours. It's a game of trial and error.

[VeryBumpy 272]

Well pure frame rate is one thing but there are also fluidity and visual anomalies that the tweaked cfg's help to eliminate.

[Bartbear 10]

Nice test, so i tried something simular but with a different approach.

 

1. I took my current scrupulous Word Not Allowed-tuned FSX.cfg and removed or disabled all Word Not Allowed advices and kept HIGHMEMFIX only.

 

So: Removed/Disabled:

 

No more ForceWindowedVsync

TEXTURE_MAX_LOAD to default 1024

FIBER_FRAME_TIME_FRACTION to default 0.33

No Usepools=0

No Affinity mask=14

LOD_RADIUS to default 4.500000.

 

Kept:

 

Framerate locked inside to 30

HIGHMEMFIX=1

 

> With NvidiaInspector: no changes: kept the same settings (See Word Not Allowed's site)

 

2. I let FSX make a new FSX.cfg and compared it with my now Word Not Allowed-tunings-free one to be sure that no other relevant things were changed.

 

So now i had the exact same situation and FSX settings in both files.

 

3. Now i did fly my ususal testscenario:

 

My Carenado Beech Baron 58: flying low and close along a steep hill in the valley above the Lugano lake in Switzerland, in side SPOTview, a place were my 29-30 frames usually drop to 27-28.

 

Testresult after flying 5 times this scenario with each FSX.cfg:

 

Using FSX.cfg with all the Word Not Allowed tuning: small drop to: 27-28 and a little microstutter.

 

Using same FSX.cfg without the Word Not Allowed tunings: drop to 22-23 and much more stutter.

 

So i stay with the Word Not Allowed settings for my rig, cause i never found any better !

 

http://#####...hardware-guide/

 

 

Intel I5 2500K overclocked @ 4.3

8 GB RAM

GTX 570 DCII 1.28 GB

ASUS P8Z68-V PRO

Intel SSD-320 160 Gb (for Win 7 and FSX)

WD Caviar Black 1 Tb

Windows 7 Home Premium SP1 - 64 bit

 

Flight Environment X, Ultimate Terrain X Europe, Ground Environment X Europe

[Guest Raul_Hernandez]

Well pure frame rate is one thing but there are also fluidity and visual anomalies that the tweaked cfg's help to eliminate.

 

True. But in my case this morning, as I said above, there is no improvement / degradation between visual quality with the tweaked and untweaked CFGs.

 

@ Bartbear ...... Interesting test too. :smile: .... I suspect most users with nVidia will see better results with a tweaked CFG... at least tweaked knowledgeably, in some areas.

 

But ATI...... is a different story I reckon.

[wanabflyer 53]

Depends on aircraft used, doesn't it?

I think it does certainly depend on the aircraft used.... (iFly 737 for FSX was one of the only products I ever got a refund on from Flight 1, due to frame rates dropping from 30 to 8, when the mouse was in view)....

 

But also the last rig that I had a few 'mouse cursor in view' issues with, was using an old 8800 GTX card.... since getting my current rig with an ATI 5770 1GB I have never seen any mouse/FPS issues, on any aircraft.

 

Perhaps it is one area where nVidia has the issue rather than ATI ... (makes a nice change! :lol:)

 

At the moment i am only flying the PMDG 737NGX. Having said that, my card is no slouch.