Who ever gets a 4gb 670 ftw won't give it away too fast. Especially if he has more than 1. The FTW version is boosted by adding more power components to the card. I also heard evga is picking up certain 670 chips for this specific version.

I've tried every combination of FSX.cfg settings as well.. these settings seem ok in general with little clouds.. sharing below.. bufferpools makes no difference at least with what i've tried.. on or off.. i've even overclocked the video card by +60 on both clocks with only about 2 fps increase roughly.

 

 

Before you spend any money on another video card I would try dumping some of those NV Inspector settings.

 

I am running something very similar to you:

3x24" screens at 6040x1200 in surround mode with a 2GB GTX670, 4770K @4.4 with HT on.

 

With DX10 I find the only thing you really need to change from default is Sparse Grid Supersampling.

I use the 4x setting as you show. This gets rid of about 95% of the shimmers.

Set Sparse Grid Supersampling to 4x but do not have any override set.

Set Antialiasing and Anisotropic filtering both back to application controlled and OFF.

This includes setting the Anisotropic Filtering Setting to OFF, not 16x.

 

Turn on Antialiasing and Anisotropic Filtering back in FSX only.

 

I use 8x CSAA in Steve's Fixer.

 

I would also start a new FSX cfg and just add the basic Hi Memory, Affinity Mask and Wide View = True changes.

 

gb.

Before you spend any money on another video card I would try dumping some of those NV Inspector settings.

 

I am running something very similar to you:

3x24" screens at 6040x1200 in surround mode with a 2GB GTX670, 4770K @4.4 with HT on.

 

With DX10 I find the only thing you really need to change from default is Sparse Grid Supersampling.

I use the 4x setting as you show. This gets rid of about 95% of the shimmers.

Set Sparse Grid Supersampling to 4x but do not have any override set.

Set Antialiasing and Anisotropic filtering both back to application controlled and OFF.

This includes setting the Anisotropic Filtering Setting to OFF, not 16x.

 

Turn on Antialiasing and Anisotropic Filtering back in FSX only.

 

I use 8x CSAA in Steve's Fixer.

 

I would also start a new FSX cfg and just add the basic Hi Memory, Affinity Mask and Wide View = True changes.

 

gb.

 

Thanks for these..

 

I tried setting 8x CSAA and fixer and the way you have things with 4xSGSS instead of 4xSupersampling..

 

I didnt tweak my fsx.cfg (start from scratch).. at least not yet.. kept what i thought were all my optimal settings..

 

This all with my full blown install.. doing Word Not Allowed's Test - Clouds AA saved flight for 60 seconds in fraps:

 

Your settings (and my old fsx.cfg):  7.43 fps

With my settings and using 4x and 0 in fsx.cfg and 4xsupersampling:  19.63

 

Ill try the fsx.cfg route one more time..

 

I'll even try this on a more "bare" system image and see (i'm sure it will go up with less scenery, like not having ftx, utx, etc).. but will still give it a shot..

 

I have till monday till the 780 Superclock by EVGA comes that i could turn it down if its all in the settings.

Your settings (and my old fsx.cfg): 7.43 fps

 

 

Hmmm, that's not good.

I am getting about 14 FPS with Word Not Allowed's Test - Clouds AA just watching the FSX frame rate display.

I might have to try your settings  

 

gb.

Hmmm, that's not good.

I am getting about 14 FPS with Word Not Allowed's Test - Clouds AA just watching the FSX frame rate display.

I might have to try your settings  

 

gb.

 

Wow double.. yeah i need to rename fsx.cfg and see how goes too..

 

Yours is in line with Kostas results from his wordpress site (cant paste link i guess)...

 

Though i dont know what resolution he was using.. You and I run very similar resolutions.

 

His charts show with that 780 card.. would gain about 6 fps possibly.. 6 fps, worth spending the extra $200 i dunno (from selling old).. maybe.. titan too expensive now.. that one appears to be the holy grail.

 

 

His charts show with that 780 card.. would gain about 6 fps possibly.. 6 fps, worth spending the extra $200 i dunno (from selling old).. maybe.. titan too expensive now.. that one appears to be the holy grail.

 

There is one post in the Monitor/Video forum saying the new GTX780ti will do better in FSX than the Titan.

 

http://forum.avsim.net/topic/424991-nvida-gtx-780-ti-specs-full-gk110-unlocked/

 

Might be worth waiting for?

 

gb.

Ah nice. Possibly. Though $200 more than a 780 at $700. In fact. The card was just released today. Tempted to get my hands on one and run some benchmarks . http://www.newegg.com/Product/Product.aspx?Item=14-487-002

 

 

Sent from my iPhone using Tapatalk

So i have updated results..

 

Using Kostas test:

 

On a clean machine I averaged 8.6 fps over 60 seconds in fraps.

 

This is in DX10 mode using the settings below..

 

 

[DISPLAY.Device.NVIDIA GeForce GTX 670.0]

Mode=5760x1200x32

Anisotropic=1

AntiAlias=1

 

[DISPLAY.Device.NVIDIA GeForce GTX 670.0.0]

AntiAlias=1

Mode=5760x1200x32

Anisotropic=1

AntiAlias=1

 

//5760x1080 or 5040x1050 or 4800x1200 or 4800 x 900 or 4320x900 (did try) or 3840x1024 or 3840 x 800 or 3840 x 768

 

[GRAPHICS]

AC_SELF_SHADOW=0

AIRCRAFT_REFLECTIONS=1

AIRCRAFT_SHADOWS=1

ALLOW_SHADER_30=1

COCKPIT_HIGH_LOD=1

D3D10=0

EFFECTS_QUALITY=2

ForceWindowedVSync=0

GROUND_SHADOWS=0

HIGHMEMFIX=1 

IMAGE_QUALITY=0

LANDING_LIGHTS=1

 

 

MultiSamplesPerPixel=4  //8 and 8 with 4x supersampling was OK

MultiSampleQuality=0

 

NUM_LIGHTS=8

See_Self=1

TEXTURE_MAX_LOAD=4096

TEXTURE_QUALITY=3

Text_Scroll=1

SHADER_CACHE_PRIMED_10=1693500672

SHADER_CACHE_PRIMED=1693500672

 

 

[TERRAIN]

AUTOGEN_DENSITY=3

BATCH_AUTOGEN2=1

DETAIL_TEXTURE=1

LOD_RADIUS=7.500000

MESH_COMPLEXITY=90

MESH_RESOLUTION=23

SWAP_WAIT_TIMEOUT=2  

//TERRAIN_MAX_AUTOGEN_BUILDINGS_PER_CELL=897     //897

//TERRAIN_MAX_AUTOGEN_TREES_PER_CELL=1100  //1100

TEXTURE_RESOLUTION=29

WATER_EFFECTS=4

IMAGE_PIXELS_FOR_AUTOGEN_POLYGONS=2048  //new

 

 

[sOUND]

AmbientUIMusicVolume=-6.000000

PrimaryDevice={00000000-0000-0000-0000-000000000000}

VoiceDevice={00000000-0000-0000-0000-000000000000}

SOUND=1

SOUND_QUALITY=2

SOUND_LOD=0

UISound=1

AmbientUI=1

 

[sCENERY]

ObjectsToBatchPerFrame=50 //new

DAWN_DUSK_SMOOTHING=1

IMAGE_COMPLEXITY=4

LENSFLARE=1

MAX_ASYNC_BATCHING_JOBS=3

SmallPartRejectRadius=1  //4 b4

 

[Display]

ActiveWindowTitleBackGroundColor=0,28,140,64

ActiveWindowTitleTextColor=255,255,255

BLOOM_EFFECTS=0

ChangeTime=4.000000

InfoBrakesEnable=True

InfoLowerLeftBackGroundColor=255,0,0,128

InfoLowerLeftTextColor=255,255,255

InfoLowerRightBackGroundColor=255,0,0,128

InfoLowerRightTextColor=255,255,255

InfoOverspeedEnable=True

InfoParkingBrakesEnable=True

InfoPauseEnable=True

InfoSlewEnable=True

InfoStallEnable=True

InfoUpperRightBackGroundColor=0,0,0,0

InfoUpperRightTextColor=255,0,0

NonActiveWindowTitleBackGroundColor=24,33,87,64

NonActiveWindowTitleTextColor=255,255,255

RUNWAY_LIGHTS_APPROACH_SCALAR=0.35

RUNWAY_LIGHTS_STROBE_SCALAR=0.35

RUNWAY_LIGHTS_SURFACE_SCALAR=0.35

RUNWAY_LIGHTS_VASI_SCALAR=0.25

SKINNED_ANIMATIONS=1

TEXTURE_BANDWIDTH_MULT=120

TextureMaxLoad=6

TransitionTime=4.000000

UPPER_FRAMERATE_LIMIT=30

WideViewAspect=True

 

[JOBSCHEDULER]

AffinityMask=84

 

[bUFFERPOOLS]

//RejectThreshold=126976

//UsePools=1

UsePools=0

//Poolsize=0

 

 

[Main]

User Objects=Airplane, Helicopter

SimObjectPaths.0=SimObjects\Airplanes

SimObjectPaths.1=SimObjects\Rotorcraft

SimObjectPaths.2=SimObjects\GroundVehicles

SimObjectPaths.3=SimObjects\Boats

SimObjectPaths.4=SimObjects\Animals

SimObjectPaths.5=SimObjects\Misc

DisablePreload=1

FIBER_FRAME_TIME_FRACTION=0.12 //0.05 

HideMenuFullscreen=1

HideMenuNormal=0

Location=382,753,6142,1953,\\.\DISPLAY1

Maximized=1

PerfBucket=7

//ProcSpeed=9788

//ProcSpeed=9593

//ProcSpeed=9453

ProcSpeed=17424

 

[PANELS]

IMAGE_QUALITY=1

QUICKTIPS=0

PANEL_OPACITY=100

PANEL_MASKING=1

PANEL_STRETCHING=1

UNITS_OF_MEASURE=0

DEFAULT_VIEW=0

 

[Weather]

WindshieldPrecipitationEffects=1

MinGustTime=10

MaxGustTime=500

MinGustRampSpeed=1

MaxGustRampSpeed=200

MinVarTime=5

MaxVarTime=50

MinVarRampSpeed=10

MaxVarRampSpeed=75

TurbulenceScale=1.000000

WeatherServerAddress=fs2k.zone.com

WeatherServerPort=80

WeatherGraphDataInDialog=0

AdjustForMagVarInDialog=1

DynamicWeather=0

DownloadWindsAloft=0

DisableTurbulence=0

CLOUD_DRAW_DISTANCE=3

DETAILED_CLOUDS=1

CLOUD_COVERAGE_DENSITY=8

THERMAL_VISUALS=0

 

[AccelerationPack]

ControlsFirstRun=0

HomePageFirstRun=0

[CONTROLS]

Controls_Default=Standard

Controls_Current=Standard

KBDAIL=64

KBDELEV=64

KBDRUD=64

[TextInfo.1]

Latitude=1,1

Longitude=1,2

Altitude=1,3

Heading=1,4

AirSpeed=1,5

WindDirectionAndSpeed=1,6

[TextInfo.2]

FrameRate=1,1

LockedFrameRate=1,2

GForce=1,3

FuelPercentage=1,4

[TextInfo.3]

Latitude=1,1

Longitude=1,2

Altitude=1,3

Heading=1,4

AirSpeed=1,5

WindDirectionAndSpeed=1,6

FrameRate=2,1

LockedFrameRate=2,2

GForce=2,3

FuelPercentage=2,4

[slewTextInfo.1]

Latitude=1,1

Longitude=1,2

Altitude=1,3

Heading=1,4

AirSpeed=1,5

[slewTextInfo.2]

FrameRate=1,1

LockedFrameRate=1,2

[slewTextInfo.3]

Latitude=1,1

Longitude=1,2

Altitude=1,3

Heading=1,4

AirSpeed=1,5

FrameRate=2,1

LockedFrameRate=2,2

[DynamicHeadMovement]

LonAccelOnHeadLon=-0.020000

LonAccelOnHeadPitch=-0.010000

RollAccelOnHeadLat=0.010000

YawAccelOnHeadLat=-0.100000

RollAccelOnHeadRoll=0.100000

MaxHeadAngle=5.000000

MaxHeadOffset=0.300000

HeadMoveTimeConstant=1.000000

[VirtualCopilot]

VirtualCopilotActive=1

[uSERINTERFACE]

PageID=6

SHOW_MISSION_CAPTIONS=0

PAUSE_ON_LOST_FOCUS=1

PROMPT_ON_EXIT=1

SITUATION=FLIGHTS\OTHER\FLTSIM

Map_Orientation=2

ShowAllACPaintSchemes=0

SelectAircraftManufacturer=Cessna

SelectAircraftPublisher=All

SelectAircraftType=All

[ATC]

ShowATCText=1

COMM_MSG_NONE_COLOR=FFFFFFFF

COMM_MSG_ATC_USER_COLOR=FFB6FFB6

COMM_MSG_USER_ATC_COLOR=FFFFD21B

COMM_MSG_ATC_AI_COLOR=FF00FF00

COMM_MSG_AI_ATC_COLOR=FFFF7840

AutoOpenAirTrafficWindow=1

UsePilotVoice=1

PilotVoice=0

[PointOfInterestSystem]

CycleSetting=0

 

 

 

 

[TrafficManager]

AirlineDensity=0

GADensity=0

FreewayDensity=7

ShipsAndFerriesDensity=0

LeisureBoatsDensity=0

IFROnly=0

AIRPORT_SCENERY_DENSITY=1

 

 

 

[AContain]

ShowLabels=1

ShowUserLabel=0

ShowLabelManufacturer=0

ShowLabelModel=1

ShowLabelTailNumber=0

ShowLabelDistance=1

ShowLabelAltitude=1

ShowLabelAirline=0

ShowLabelAirlineAndFlightNumber=0

ShowLabelFlightPlan=0

ShowLabelContainerId=0

ShowLabelAirspeed=0

ShowLabelHeading=0

LabelDelay=1000

LabelColor=FFFF0000

[iNTERNATIONAL]

ASLAT=2

ASLON=1

MEASURE=0

[REALISM]

PFactor=0.000000

Torque=0.000000

GyroEffect=0.000000

CrashTolerance=0.000000

General=0.000000

UnlimitedFuel=True

AllowEngineDamage=False

TrueAirspeed=False

AutoCoord=False

RealMixture=False

StressDamage=False

GEffect=False

ManualLights=False

GyroDrift=False

CrashWithDyn=False

CrashDetection=False

AutoTrim=False

[sIM]

SYSCLOCK=1

[sTARTUP]

DEMO=0

SHOW_OPENING_SCREEN=1

STARTUP_DEMO=

LoadWindow=1

[FACILITIES]

COUNTRY=

STATE=

CITY=

GTL_BUTTON=4096

[Misc]

Com_Rate=7

 

 

[DISPLAY.Device.NVIDIA GeForce GTX 670.0.0]

Mode=5760x1200x32

Anisotropic=1

AntiAlias=1

[MULTIPLAYER]

condAccoutPassword=0

 

I then decided to switch to dx9, just for fun..

 

I averaged 16.3 fps under dx9.. amazing.. i had sworn that dx10 was supposed to give better performance than dx9?

 

In both cases I kept the NI settings the same as below (my original test on my fully installed fsx install was around 7.5 fps.. so about 1 fps better on this "clean image".. not worth redoing but a little better.. again on the EVGA 670 FTW 2GB card at 4.6 ghz on an i7 2600k

 

 

 

I also did comparisons between AS2012 and REX.. rex was about 0.5 fps worse when using 2048.. 

 

2048 vs 4096 i could see no visual difference in switching back and forth in windows photo viewer.. there was no real performance loss by leaving 4096 on either.  I did try 32bit water and other texture settings instead of dxt1.. that was a killer..

 

The test avg became 3.08 with those enabled.

"The test avg became 3.08 with those enabled."

 

Just as well you are overclocked.

 

gb.

Found something interesting tonight.. in my dll.xml there are many entries that have contributed to about 8 fps loss in performance.. some of which are a few objectflow entries for ftx.. still narrowing things down.. but basically my one local airport test aiming at the skies is now near 30 fps with full on clouds where before it was much lower (4xSS not sgss).. still some merit for a 780 or 780ti for heavy weather.. ksea still sinks to 18-24 absolute worst in stormy weather.. still not bad though.

 

I have both cards on the way, 780ti and 780.. one is not going to be staying for sure.. leaning towards the TI staying:)'

 

I do need to rerun the AA cloud test again with the "clean" dll.xml.. and report that retest fps (doubt it will go up, as the above aa tests were done on a clean partition).

Thanks for these..

 

I tried setting 8x CSAA and fixer and the way you have things with 4xSGSS instead of 4xSupersampling..

 

I didnt tweak my fsx.cfg (start from scratch).. at least not yet.. kept what i thought were all my optimal settings..

 

This all with my full blown install.. doing Word Not Allowed's Test - Clouds AA saved flight for 60 seconds in fraps:

 

Your settings (and my old fsx.cfg):  7.43 fps

With my settings and using 4x and 0 in fsx.cfg and 4xsupersampling:  19.63

 

Ill try the fsx.cfg route one more time..

 

I'll even try this on a more "bare" system image and see (i'm sure it will go up with less scenery, like not having ftx, utx, etc).. but will still give it a shot..

 

I have till monday till the 780 Superclock by EVGA comes that i could turn it down if its all in the settings.

 

Updated testing.. on my win 8.1 machine..

 

Still clocking in at 8.5 fps with 4xSGSS on the cloud test.. (7.5 lod, 2048 textures combo of rexe and as2012)

 

However.. getting around 20.8 with 4xSupersample and 4xFixer, which i hadnt tried before.. This is in DX10 mode. (7.5 lod)

 

In DX9 mode its 17.7 fps with 4Xsupersample.. so this is the usual DX10 mode gives better fps that I should have gotten before.

 

Now the curious test will be to see how the EVGA 780ti compares.. i hope to get 24.5 fps for the 4xSGSS instead of 8.5.. in which case this will be golden.

 

Again these are with the fsx.cfg settings from my original post with those settings (cars at 7% etc) and 4x in Fixer (samplesperpixel 4 and quality 0)

 

I think the take home here is.. that on a full blown fsx install, if fps seem way off.. start with a bare dll.xml file to rule out weirdness there.. then work on fsx.cfg and NI settings after that has been ruled out.

 

Incidentally, at least with 4xSupersample.. switching from 4x to 8xCSAA made no change in fps which is good.

I have some initial results from my 780Ti.. i've posted them here in this thread:  http://forum.avsim.net/topic/426053-my-experiences-with-the-evga-780ti-3gb-vs-670ftw-and-5760x1200/

I've been toying with the 4770k + 780ti now for over a month.. interesting find i found was that even at near stock speeds.. the 4770k haswell cpu showed ZERO fps decrease compared to say 4.5 ghz.. this using the same Word Not Allowed AA cloud test sequence over 60 seconds.. still around 12-13 fps.. and so much for a 10% boost.. at least with THIS test (from sandy bridge i7-2600k 4.6 that i was at).. other flight situations may show that 10% cpu boost from where i was.. or perhaps just the combo of a killer gpu with an already decent cpu is why overclocking or switching from SB to Haswell showed no real change with this particular test.. more testing needed though.