As part of the never ending tuning process of FSX I've been getting some interesting results that really make no sense. I did some extensive testing last night that has convinced me that the tweaking process for FSX is like riding the Crazy Train non-stop service to Cuckooville!For each scenario below. Scenery is at VERY dense and Autogen is at Dense. Water is at 2x Low, Cloud Draw distance is at 90 miles. 1m resolution. No aircraft shadows. Commercial traffic at 100%, General Aviation at 30% and other AI at the defaults. 1920x1280 Resolution. 4x AA (Box), AF 16x, Mipmap High Quality, Adaptive AA. Absolutely no blurries, stuttering or any other issues. FSX looks stunning.The purposes of the tests outlined below is only to see if it is possible to get more FPS but not to troubleshoot any issues.System Spec .QX9650 @ 4.02GhzATI 4890 OC - Cat 10.124 GB OCZ Platinum Ram 9-9-9-20 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test Scenario:1. Friday Island ORBX PNW Scenery2. Trike aircraft3. REX HD Clouds - Stormy4. Bufferpools 0, Fiber_Frames=0.335. Unlimited Frame Rate.With the above mentioned configuration I am getting an average of 32 FPS.NextSystem Spec QX9650 @ 4.02GhzATI 5970 (no overclock) Cat 10.12 (Notice here the architecture change to a lower clocked multi-GPU with more graphic processing capabilities)8GB OCZ Dominator 8-9-9-18 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test ScenarioSame as aboveWith the above mentioned configuration I am getting an average of 30 FPS.Next.QX9650 @ 3.0 Ghz (no overclock)ATI 5970 8GB OCZ Dominator 9-9-9-20 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test ScenarioSame as aboveWith the above mentioned configuration I am getting an average of 32 FPS.Here are my conclusions after a few weeks of testing.1. What this shows above is the addition of the ATI 5970 vs the ATI 4890 has absolutely no effect on FSX despite the fact that performance between the two cards are stunningly different. Where the 5970 shows scores 2x or 3x that of the 4890 in synthetic benchmark tests and is still considered the world's fastest GPU (arguably), FSX has reached the limit as to what it can pump through the GPU.2. Upgrading from 4GB to 8GB has absolutely no effect on FSX. Similarly the timings and overclock for the memory have almost no effect. Moving from Platinum to Dominator at 1333Mhz provides no enhancement in FSX. 3. There is no appreciable difference in FSX performance between 3GHz and 4GHz on the QX9650. One might have assumed that FSX is CPU bound so the biggest performance gain would be to increase the CPU clock but that seems not to be the case. It seems that for a given CPU architecture that FSX will eventually hit a wall at which no hardware speed increases will have an effect.4. *******' tweaks (with the exception of Buffer Pools=0) really have little effect on this configuration. As a matter of fact a line-by-line tweak copy of *******' tweaks actually yields LOWER FPS in FSX. Additionally the Shader 3.0 file also slows down the configuration or has absolutely no effect either on the 4890 or 5970 GPUs.Now. If I test other games such as popular FPS or MMORPG the difference in frame rates is hugely different. In games such as Crysis the difference between video cards is on the order of magnitude of 2 or better. Same thing with a 1Ghz overclock on the CPU that alone affects FPS in some cases by 20 - 50 FPS.So what does that tell you?IF you know what you are doing and have a properly optimized system further tweaking is going to yield VERY limited results in FSX. Accept what you have and move on. FSX seems very platform dependent so endlessly following new tweaks for a given platform is fools errand. Putting in more RAM won't help you if you are already at 4GB BUT if you have a lower class CPU within your supported Chipset, a CPU upgrade along with faster RAM will help add some FPS to FSX. Once you reach the top for your given platform you're done. In the case of the QX9650 and the Maximus Extreme MB there is simply no where else to go with that platform and all the tweaking in the world won't change that.Mind you, my conclusions ONLY apply to FSX. If you also play other games you will see huge performance gains as you beef up your rig. FYI, The last computer I purchased was a 486DX2-66 so I have been hand-building and tweaking my rigs for more than 12 years so I'm no system novice. The moral of the story is to optimize your system irrespective of FSX at which point FSX tweaks simply become irrelevant except to smooth out the game. In the same token, endlessly following tweaks is the FSX Crazy Train. This can be validated by simmers with Intel 980X powered rigs poking around here trying to find tweaks because of "poor" performance. This should come as no surprise since going from a 4 core to an 8 core at the same or lower processing speed is not advantageous to FSX. Likewise going from a single GPU at a higher clock to a muti-GPU at the same or higher clock (or even lower clock) doesn't impress FSX...it simply can't and doesn't take advantage of the newest technologies. The crazy part is that MOST other games made at around the same time at FSX do indeed take advantage of new technologies. For instance: I've seen WOW go from ~150 FPS to over 300 FPS with the same configs above. Of course you don't need 300 fps in FSX BUT it shows how the game's performance shifts upward as you upgrade. Soooo, I'm waiting for the Sandybridge CPUs to be released later this year and will build a new rig. However, since my rig is still blazingly fast and can run any of the latest releases on high settings I will probably wait until Flight is released before I do so. If Flight is coded like FSX spending $4000 to get another 10 FPS is just not worth it so we'll just have to wait and see what comes down the pike. I'm officially off the Crazy Train and am simply going to enjoy FSX at its current level and accept the fact: "what is...is"Just some interesting observations I thought that I would share...your mileage can and WILL vary.

Also a recent article in TH shows that no games benefit from 4 to 8 to 16 GB of memory. Right now the sweet spot is still 4GB even running a 64bit OS.

the only tweaks you need are the sliders.... getting down and dirty into finding some majic number is useless. You want relative numbers 20% 40% 60% not 439820595 or whatever.

As part of the never ending tuning process of FSX I've been getting some interesting results that really make no sense. I did some extensive testing last night that has convinced me that the tweaking process for FSX is like riding the Crazy Train non-stop service to Cuckooville!For each scenario below. Scenery is at VERY dense and Autogen is at Dense. Water is at 2x Low, Cloud Draw distance is at 90 miles. 1m resolution. No aircraft shadows. Commercial traffic at 100%, General Aviation at 30% and other AI at the defaults. 1920x1280 Resolution. 4x AA (Box), AF 16x, Mipmap High Quality, Adaptive AA. Absolutely no blurries, stuttering or any other issues. FSX looks stunning.The purposes of the tests outlined below is only to see if it is possible to get more FPS but not to troubleshoot any issues.System Spec .QX9650 @ 4.02GhzATI 4890 OC - Cat 10.124 GB OCZ Platinum Ram 9-9-9-20 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test Scenario:1. Friday Island ORBX PNW Scenery2. Trike aircraft3. REX HD Clouds - Stormy4. Bufferpools 0, Fiber_Frames=0.335. Unlimited Frame Rate.With the above mentioned configuration I am getting an average of 32 FPS.NextSystem Spec QX9650 @ 4.02GhzATI 5970 (no overclock) Cat 10.12 (Notice here the architecture change to a lower clocked multi-GPU with more graphic processing capabilities)8GB OCZ Dominator 8-9-9-18 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test ScenarioSame as aboveWith the above mentioned configuration I am getting an average of 30 FPS.Next.QX9650 @ 3.0 Ghz (no overclock)ATI 5970 8GB OCZ Dominator 9-9-9-20 2TWindows 7 CPU score = 7.4Windows 7 Memory score = 7.4Test ScenarioSame as aboveWith the above mentioned configuration I am getting an average of 32 FPS.Here are my conclusions after a few weeks of testing.1. What this shows above is the addition of the ATI 5970 vs the ATI 4890 has absolutely no effect on FSX despite the fact that performance between the two cards are stunningly different. Where the 5970 shows scores 2x or 3x that of the 4890 in synthetic benchmark tests and is still considered the world's fastest GPU (arguably), FSX has reached the limit as to what it can pump through the GPU.2. Upgrading from 4GB to 8GB has absolutely no effect on FSX. Similarly the timings and overclock for the memory have almost no effect. Moving from Platinum to Dominator at 1333Mhz provides no enhancement in FSX. 3. There is no appreciable difference in FSX performance between 3GHz and 4GHz on the QX9650. One might have assumed that FSX is CPU bound so the biggest performance gain would be to increase the CPU clock but that seems not to be the case. It seems that for a given CPU architecture that FSX will eventually hit a wall at which no hardware speed increases will have an effect.4. *******' tweaks (with the exception of Buffer Pools=0) really have little effect on this configuration. As a matter of fact a line-by-line tweak copy of *******' tweaks actually yields LOWER FPS in FSX. Additionally the Shader 3.0 file also slows down the configuration or has absolutely no effect either on the 4890 or 5970 GPUs.Now. If I test other games such as popular FPS or MMORPG the difference in frame rates is hugely different. In games such as Crysis the difference between video cards is on the order of magnitude of 2 or better. Same thing with a 1Ghz overclock on the CPU that alone affects FPS in some cases by 20 - 50 FPS.So what does that tell you?IF you know what you are doing and have a properly optimized system further tweaking is going to yield VERY limited results in FSX. Accept what you have and move on. FSX seems very platform dependent so endlessly following new tweaks for a given platform is fools errand. Putting in more RAM won't help you if you are already at 4GB BUT if you have a lower class CPU within your supported Chipset, a CPU upgrade along with faster RAM will help add some FPS to FSX. Once you reach the top for your given platform you're done. In the case of the QX9650 and the Maximus Extreme MB there is simply no where else to go with that platform and all the tweaking in the world won't change that.Mind you, my conclusions ONLY apply to FSX. If you also play other games you will see huge performance gains as you beef up your rig. FYI, The last computer I purchased was a 486DX2-66 so I have been hand-building and tweaking my rigs for more than 12 years so I'm no system novice. The moral of the story is to optimize your system irrespective of FSX at which point FSX tweaks simply become irrelevant except to smooth out the game. In the same token, endlessly following tweaks is the FSX Crazy Train. This can be validated by simmers with Intel 980X powered rigs poking around here trying to find tweaks because of "poor" performance. This should come as no surprise since going from a 4 core to an 8 core at the same or lower processing speed is not advantageous to FSX. Likewise going from a single GPU at a higher clock to a muti-GPU at the same or higher clock (or even lower clock) doesn't impress FSX...it simply can't and doesn't take advantage of the newest technologies. The crazy part is that MOST other games made at around the same time at FSX do indeed take advantage of new technologies. For instance: I've seen WOW go from ~150 FPS to over 300 FPS with the same configs above. Of course you don't need 300 fps in FSX BUT it shows how the game's performance shifts upward as you upgrade. Soooo, I'm waiting for the Sandybridge CPUs to be released later this year and will build a new rig. However, since my rig is still blazingly fast and can run any of the latest releases on high settings I will probably wait until Flight is released before I do so. If Flight is coded like FSX spending $4000 to get another 10 FPS is just not worth it so we'll just have to wait and see what comes down the pike. I'm officially off the Crazy Train and am simply going to enjoy FSX at its current level and accept the fact: "what is...is"Just some interesting observations I thought that I would share...your mileage can and WILL vary.

Excellent observations and a well reasoned approach. I suspect a fair number of folks with capable modern systems would be much happier just leaving things alone.DJ

I agree that FSX's performance is system dependent, however fsc.cfg tweaks have had dramatic effects on my system.The right bufferpools setting has increased fps and created a much smoother sim.FFTF has improved texture loading and sharper images.There are many more, but I'll conclude with, tweaking works for me.

The only thing your test really shows is that there is some kind of bottleneck somewhere that is holding performance back. Exactly where that bottleneck is is anyones guess

Mike T,Excellent post :( I've spent som time testing and benchmarking several of the exotic tweaks. None of them gave any increased IQ or performance. Waste of time EDITED:I do use the following non-exotic tweaks:

TEXTURE_BANDWIDTH_MULT=100Poolsize=90000000

And these settings, which are not performance related tweaks:

HIGHMEMFIX=1TEXTURE_MAX_LOAD=4096LOD_RADIUS=5.500000

Hi Mike,Great post!! :( I fully agree with you. I've stopped tweaking because it no longer improves anything on my old system. In fact I prefer flying.One of the best tweak I have is to stop looking at the FPS counter. Everybody should try it, they will be impressed by the result. :(

It is amazing how FSX confuses us about performance - no two systems are alike. I have experienced improvements by the Needham/Bojote tweaks and just Scenery Slider settings to get a fast(60fps)/smooth system with a Q6600 @ 2.7Ghz (with 800Mhz memory) but when I tried to use that system with three windowed mode, independent views (210 degree field of view total) in a TH2Go configuration, I ended up getting 15->20 fps. When I tried the same exact test on a [email protected] with a NV470 video card, I get 40->50 fps. The 980X with QPI memory management and CL7 1600Mhz memory really has some FSX power. Wish it was a little more affordable. I do think that memory speed ratios can predict a significant performance gain for FSX.

Also a recent article in TH shows that no games benefit from 4 to 8 to 16 GB of memory. Right now the sweet spot is still 4GB even running a 64bit OS.

Have to respectfully disagree with you there. I've gots tons of games and sims installed, and though that still may be true for most of those games, my installation of FSX routinely uses at least 4gb of RAM. During session I'm often looking at between 4gb and 5.5gb of RAM being actively used, with the remaining 4 - 3 gigs cached (out of 8gb available). At one point the RAM usage peaked out at 6.2gb!There are several reasons for this.Firstly, sitting idle my system is already eating up 1.2 to 1.5 gigabytes of RAM. However, one of the positive trade-offs is that many installed apps start very quickly, including FSX.Second, my system specs:> Homebuilt> Intel i5 750 OC'ed to 3.3GHz, w/CoolerMaster Hyper N520> 8GB of Corsair 1600 RAM> ATI HD 4770 GPU (best "bang for the buck" graphics card I've ever had!)> ASUS P7P55D LE> Windows 7, 64 bit Third, like many of us I have a highly stoked-up, tweaked-out fsx.cfg. Notable inclusions are:[bufferPools]RejectThreshold=98304[DISPLAY.Device.ATI Radeon HD 4770.0.0]Anisotropic=1AntiAlias=1Mode=1680x1050x32[GRAPHICS]AC_SELF_SHADOW=1AIRCRAFT_REFLECTIONS=1AIRCRAFT_SHADOWS=1ALLOW_SHADER_30=1COCKPIT_HIGH_LOD=1D3D10=0EFFECTS_QUALITY=2ForceFullScreenVSync=0ForceWindowedVSync=1GROUND_SHADOWS=0HIGHMEMFIX=1IMAGE_QUALITY=0LANDING_LIGHTS=1NUM_LIGHTS=8STALE_BUFFER_THRESHOLD=2147483647See_Self=1TEXTURE_MAX_LOAD=2048TEXTURE_QUALITY=3Text_Scroll=0SHADER_CACHE_PRIMED_10=1693500672SHADER_CACHE_PRIMED=1693500672[Display]...BLOOM_EFFECTS=1...RUNWAY_LIGHTS_APPROACH_SCALAR=0.3RUNWAY_LIGHTS_STROBE_SCALAR=0.3RUNWAY_LIGHTS_SURFACE_SCALAR=0.3RUNWAY_LIGHTS_TAXI_SCALAR=0.3RUNWAY_LIGHTS_VASI_SCALAR=0.3SKINNED_ANIMATIONS=1TEXTURE_BANDWIDTH_MULT=90TextureMaxLoad=90TransitionTime=4.000000UPPER_FRAMERATE_LIMIT=25WideViewAspect=True[TERRAIN]AUTOGEN_DENSITY=5DETAIL_TEXTURE=1LOD_RADIUS=9.000000MESH_COMPLEXITY=100MESH_RESOLUTION=25SWAP_WAIT_TIMEOUT=10 // was 10TEXTURE_RESOLUTION=29WATER_EFFECTS=7[sCENERY]DAWN_DUSK_SMOOTHING=1IMAGE_COMPLEXITY=5LENSFLARE=1MAX_ASYNC_BATCHING_JOBS=3I have the "Job Scheduler" and "Affinity Mask" commented out. With the former enabled I saw no improvement in performance, and with the latter I actually saw a decrease in performance along with only three out of the four cores actually being used (though all four were selected!).Third: Active Sky Evolution and Ultimate Traffic II are running in the background, and I have several tens of gigabytes worth of scenery and high-res mesh installed, as well as a godawful amount of custom GA traffic. I have only a handful of add-on aircraft, though. I don't want to saturate the hangar with a plethora of aircraft, thereby diluting the value of flying any one of them, and also making it more difficult choosing an aircraft to fly.Now, most of the time I'm in the flight levels, often above FL330 and sometimes above FL400. About a third of the time I'll fly bug smashers. No matter the flight, though, lots of RAM is gobbled up. The busier the airspace, the more dramatic the weather, and the more detailed the scenery, the more RAM is being used. This RAM consumption is desired because we know that if FS calls for more RAM than is being made available, things in the FS world begin to lose textures, turn into weird "spikes", or disappear altogether, often before ultimately CTD'ing. The "HIGHMEMFIX" tag in fsx.cfg is probably the single biggest effect in stabilizing RAM usage in FSX, but this is only going to help if your system has plenty of RAM to spare (like, a minimum of 6gb plenty...), and only if you're running a 64 bit OS. Systems with 32 bit OS's don't use more than 3gb of RAM total. If you have 4gb installed on a 32bit OS system, then you have a whole gig a RAM sitting idle, never being used...The change in the .cfg that arguably has the effect of consuming the most RAM during an FSX session, and in conjunction with the other settings shown above, is increasing the LOD_RADIUS line to something greater than 4.5, which is the in-game UI allowed maximum. The maximum allowed, "manually entered" setting in fsx.cfg is 9.0. What this does is increase the area of high detail in the FSX world, both on the ground and in the air (you can spot aircraft farther away, not just the contrail with an invisible aircraft at the head of it). Fine detail textures are rendered much farther out at 9.0 than they are at 4.5. Mesh resolution at a distance is also positively affected, so you see high resolution mesh rendered much farther away at 9.0 than you would at 4.5. The results of this is a much more realistic looking world below you, even with default textures, particularly if you're flying airliners or business jets. You can see fine details in mountains 80 or more miles away. You can see roads wind and stretch until you visually can't see them anymore, or until they end. It's a huge, huge difference and it all looks very nice! However, increasing this setting is probably more taxing on your system than any other, as it's sort of a "global detail multiplier". I have mine cranked all the way because my system can handle it and the hit on performance is acceptable. Many of you have more capable systems than I do so you should try it!Also, again with the settings shown above, the "blurries" are virtually eliminated.How I have Active Sky Evolution and FSUIPC set up plays a big role in how realistic the surrounding world looks and behaves. In ASE I have visibility maxed out at 120 NM, and I DO NOT use wind and visibility smoothing (or Direct Weather Control in v642) in ASE - I use the settings in FSUIPC for that. The reason is that with ASE, if you use wind smoothing, you have to turn on Direct Weather Control, and this setting "homogenizes" the weather, i.e. you get globalized weather in the FS world. If the current station calls for a particular type of weather, then all stations will have exactly the same weather. When the weather for that one station changes, or when you fly into another weather zone and the weather changes, then the weather at all stations change. The result is that you have exactly the same weather, and the same weather changes, for as far as you can see. This might actually be desireable in some cases, but if you're up high jet-setting around the world, you might actually want to see the weather changes ahead of you. You might actually want to fly over a clear area while looking at some cumulonimbus clouds in the distance. I know I do, as this is visually as close to what is seen when flying high in reality.Setting ASE's maximum visibility to 110-120 or so is another trick that can greatly enhance the visuals of the high-altitude world. FSX's default visibility at high altitude is junk, as you will never see more than perhaps 120 or 130 miles ahead of you, and you certainly will not see the hard edge of the earth's horizon like you would in the sim with default visibility. There is simply too much atmosphere to see through.Apparently, Pete Dowson is a bit pessimistic when it comes to his confidence on how enabling the experimental "Visibility" tab in FSUIPC 4.60 will actually work in FSX. I've enabled it, made settings, and have seen a positive difference. In ASE, you can set maximum surface visibilty and maximum upper visibility, but you will still get abrupt visibility changes when traversing the layers in between. With "Visibility" enabled in FSUIPC 4 and with preferences set within that tab, I've since seen none of the abrupt changes in visibility when making changes in altitude. Visibility changes are much more consistent now. On the other hand, the rapid-but-still-gradual changes in visibility as you fly through clouds still occurs, and is desirable as it's a nice visual effect.Of course, all of this tweaking comes at a cost in performance, and this is where one must make highly subjective compromises. While taxiing at busy international airports my performance is at it's worst, as things almost become a slideshow in the lower teens, to the upper single digits at the most congested mega-airports during peak hours. At smaller, more hospitable locations I can taxi around at 15 or more fps. Dramatic weather will obviously slow things down a bit. Once I'm up flying, however, things become much smoother as the 25fps limit is frquently pegged for most of the remaining flight. I use the fps limit because I prefer enhanced visuals to outright fps performance. I could get better performance, but at the cost of the truly excellent visuals I'm enjoying now. The trick is to find a nice balance.OK, do I've digressed from the RAM topic a bit, but I hope this was still helpful. Also, if anyone has any suggestions for my .cfg, then by all means...Tree.

Have to respectfully disagree with you there. I've gots tons of games and sims installed, and though that still may be true for most of those games, my installation of FSX routinely uses at least 4gb of RAM. During session I'm often looking at between 4gb and 5.5gb of RAM being actively used, with the remaining 4 - 3 gigs cached (out of 8gb available). At one point the RAM usage peaked out at 6.2gb!...

As FSX.exe is a 32 bit application it is only capable of addressing 4 gigs of memory, so it would be technically impossible for FSX to use more than 4 gigs. The whole system might of course use more than 4 gigs, so there are some advantages of having more than 4 gigs in an FSX rig

Interestingly enough, since I was in a hardware tweaking mood I DID indeed find a pretty whopping 10 FPS boost this morning.When setting my rig back to the 4.02Ghz Overclock after all the testing: I found a setting in the BIOS that pumps up the Memory clocking in some way. I changed it from "Light" to "Moderate" ("Strong" prevents Windows 7 from booting up.) I also changed the FSX multi-core setting from 12 to 14 in the FSX.cfg. Anyway, I went back into FSX just to check to make sure I didn't break anything while doing this test - only expecting to be back where I was.WELL: In the tests above I averaged about ~32 FPS and with this one BIOS setting (maybe the 12 to 14 change too...but I don't feel like testing any more) I am now averaging ~42 FPS in the same area. Interestingly enough also, my Windows 7 CPU and Memory scores went from 7.4 to 7.5. I played around with some other clock settings but I hit a wall at ~1600Mhz FSB where Windows becomes unstable and I fail Memtest95. I am indeed seeing a 10 FPS boost in all areas so where FSX was flyable before even in the most dense areas, it now is really at movie smoothness in most areas.So anyway, I got a nice surprise on a sunny cold afternoon, now as Continental says: "It's Time to Fly!

As FSX.exe is a 32 bit application it is only capable of addressing 4 gigs of memory, so it would be technically impossible for FSX to use more than 4 gigs. The whole system might of course use more than 4 gigs, so there are some advantages of having more than 4 gigs in an FSX rig

Absolutely correct. The OS, plus UT2, plus ASE, plus FSX, is going to consume a lot of memory. I didn't mean to imply that FSX itself can use more than 4gigs (though more or less that is pretty much what I typed). Thanks for pointing that out.

What everyone forgets when looking at tweaks is the "Virtual Address Space" that FSX has to run in, ss that it can then communicate with the hardware, ie cpu physical RAM and VRAM. The limit of the VAS on a 64-bit OS is 4GB (2GB on a 32-bit OS) Running a vanilla install of FSX which has a huge number of small files and the VAS is well managed and the MS Garbage Collector works very efficiently and on a fast cpu, etc FSX runs very well, no stutters, ete etc. Now start to add on 3rd party software such as aircraft (complex) photoscenery, weather engines, etc and these now also have to be loaded into the VAS and if some of these 3rd party add-ons are large files (as opposed the the FSX small files) then they may not be released from the VAS as efficiently as the native FSX files. So the VAS becomes "saturated and bloated" and is less efficient and so the flow of data to your very fast cpu, RAM and gpu may no be longer as efficient as it could be and so FSX doesn't run as well as you expect it to. If you exceed the VAS you get the dreaded OOM message so it is logical that as the VAS "fills up" things will be processed slower. Now add in the fanatical tweaks which may or may not affect the way FSX interacts with the computer hardware and you have an electronic brew which will interact differently on every system. In FSX there is "No one size fits all"! Nor am I saying that tweaks should not be used it is just that we do not really understand what is happening. :( I am now using xperf/process explorer to try and determine what is happening in a computer when FSX is running amd what I am seeing that the number of 'event handles' FSX uses in complex areas exceeds 20,000 and that coupled with other things happening simultaneously could slow the system down. The next thing that I want to monitor is the way that FSX handles 'HEAPS' and 'STACKS' as I feel that both of these could have a significant impact on the performance of FSX. I could be totally on the wrong track but my own copy of FSX does not lose performance until I add on 3rd party software or add a tweak or two! :( What a great set of postsPeterH

I think Austin has the greatest tweak for FSX ever... his rule... VRAM is what should be used at all time, having to go to system ram for an object or texture is 10x slower, having to go to the harddrive if that object isn't in System RAM is 100x slower....A little bit exagerated but still it's close enough to make you want to tame your session... Youre not getting 20fps because the system can't handle it.. it's because your bottlenecking and the only cure is turn down the sliders or get better hardware.