Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[SAS~Storebror]

Yes of course, why not use The Black Death.
Let me summarize my specs first:
Core i5-2500K @4.6 GHz.
8 GB RAM.
Gigabyte Geforce GTX 550Ti 1GB GDDR5, GPU Clock 970 MHz, Memory Clock 2100 MHz, Shader Clock 1940 MHz.
FullHD Screen (1080p).

Ultrapack 3.0RC4, no mods.
conf.ini (irrelevant parts stripped):

Code: [Select]

[window]
width=1920
height=1080
FullScreen=1
ColourBits=32
DepthBits=24
StencilBits=8
ChangeScreenRes=0
DrawIfNotFocused=1
EnableResize=0
EnableClose=1
SaveAspect=0
Use3Renders=0
WideScreenFoV=1

[GLPROVIDER]
GL=Opengl32.dll

[game]
Arcade=0
HighGore=1

[rts]
ProcessAffinityMask=14
mouseUse=2
joyUse=1
trackIRUse=1
DisableIME=0
locale=de
JoyProfile=0
UseSmartAxis=0

[Render_OpenGL]
TexQual=3
TexMipFilter=3
TexCompress=0
TexFlags.UseDither=0
TexFlags.UseAlpha=0
TexFlags.UseIndex=0
TexFlags.PolygonStipple=0
TexFlags.UseClampedSprites=0
TexFlags.DrawLandByTriangles=0
TexFlags.UseVertexArrays=1
TexFlags.DisableAPIExtensions=0
TexFlags.ARBMultitextureExt=1
TexFlags.TexEnvCombineExt=1
TexFlags.SecondaryColorExt=1
TexFlags.VertexArrayExt=1
TexFlags.ClipHintExt=1
TexFlags.UsePaletteExt=1
TexFlags.TexAnisotropicExt=1
TexFlags.TexCompressARBExt=0

TexFlags.TexEnvCombine4NV=1
TexFlags.TexEnvCombineDot3=1
TexFlags.DepthClampNV=1
TexFlags.SeparateSpecular=1
TexFlags.TextureShaderNV=1

HardwareShaders=1

Shadows=2
Specular=2
SpecularLight=2
DiffuseLight=2
DynamicalLights=1
MeshDetail=2
VisibilityDistance=3

Sky=2
Forest=3
LandShading=3
LandDetails=2

LandGeom=2
TexLarge=1
TexLandQual=3
TexLandLarge=1

VideoSetupId=17
Water=2
Effects=1
ForceShaders1x=0
PolygonOffsetFactor=-0.15
PolygonOffsetUnits=-3.0

FogLow=2
FogHaze=4
Selector RAM setting = 512MB

Nvidia Inspector Profile:


Test result (90 seconds The Black Death) with these settings:
FPS AVG: 115 MAX: 158 MIN: 45 (1st incendiary bomb on truck column)

Changing the Inspector settings to FPS limiter @58FPS and VSYNC enabled gives this result:
FPS AVG:  56 MAX:  60 MIN: 27 (1st scene change)

Changing the Inspector settings to your 16xS AA + 8x transp. AA (no VSYNC etc., see image below) gives this result:
FPS AVG:  14 MAX:  27 MIN:  8 (at any random time) -->> unplayable.



Best regards - Mike

[rruff]

Test result (90 seconds The Black Death) with these settings:
FPS AVG: 115 MAX: 158 MIN: 45 (1st incendiary bomb on truck column)

Changing the Inspector settings to your 16xS AA + 8x transp. AA (no VSYNC etc., see image below) gives this result:
FPS AVG:  14 MAX:  27 MIN:  8 (at any random time) -->> unplayable.

Setting everything the same I get:

FPS Avg: 197 Max: 350 Min: 52 ... and
FPS Avg: 59  Max: 90   Min: 37

With the AA off my card never maxed out during the test, and the higher AA settings were clearly beyond your card's capabilities. So it's hard to tell. Wouldn't you expect less than 2x difference between our fps? How about 4x or 8x AA and maybe letting the game set AF (which is 4x, yes)? I normally just use the game's conf.ini AF set on 3,3,2.

EDIT: Just did some more tests with AF set by the application, 4x transparency SGSS, and the following AA settings:

AA setting, Avg, Max, Min FPS

8x [8x CSAA (4 color + 4 cv samples)]          183, 302, 43
4x [4x Multisampling],                                160, 259, 43
8xQ [8x Multisampling]                               114, 150, 42
8xS [Combined: 1x2 SS + 4x MS]                 107, 146, 42
8xSQ [Combined: 2x2 SS + 2x MS]               71,  109, 42

One thing I noticed was that my processor was pegged for every run. The GPU never hit 100% with the first two (8x and 4x), but by the last one (8xSQ) it was pegged nearly all the time. It only dipped when the processor was overloaded, causing the fps to drop.

It's interesting that the minimum frame rate was essentially the same for every run. Whatever is happening then (explosions and smoke) is hitting the processor independent of the AA settings.

Up to the 8xS setting I saw pretty significant improvements in appearance. The nicer settings are still noticeable, but not a big deal. 

[SAS~Storebror]

AF set by the application

That might be one of the big differences.
I remember I've tested turning off AF in Nvidia control panel once and the results were disgusting. I couldn't see any details anymore, starting from somewhere like 10m in front of my point of view.
That's why I reverted to 16xAF initially and turned it down to 8xAF later.

I'm not at my 550Ti system right now but just repeated the test with a G210 GPU (yep, horribly weak, but good for testing lowest specs).
Indeed it seems like the game's internal AF is something like 4x and when I set the same 4xAF "user controlled" in Nvidia control panel, the game looks nearly the same but FPS drops significantly.
I'll repeat that test when I'm back at my 550 system. No clue why things looked so ugly when AF was turned off there back then, but I'll check that.

Wouldn't you expect less than 2x difference between our fps?

No. Pixel rate and texture rate are 2.2x / 2.8x higher on your card and you have 640 shader units compared to my 192. That's a huge difference once you enable "eye candy" settings.

Best regards - Mike

[rruff]

Be interested to see what you find out. Just to clarify, I have the non-Ti GTX 750 with 512 shaders... still a huge difference.

I tried using the AF settings in Inspector but I got a weird stutter (even though fps was ok) and it didn't seem like it looked any better. I know that on my old card the AF settings in Inspector didn't even do anything... the game looked and performed the same regardless. It would use the conf.ini settings, though.

Don't know if you are in the market, but I've been watching prices lately (need another card for the wife's computer). GTX 660 can be had for $100 (no rebate), and GTX 750 Ti for $90 after rebate. The 660 would be a little faster but 2x the watts.

[SAS~Storebror]

I stopped watching the market now that my GTX 970 arrived.
I'm just not in the mood to install it right now, likely it will be after christmas until I find time to do so.

Best regards - Mike

[rruff]

Whoa! Nice card. You'll be looking for new ways to max that out.

[SAS~Storebror]

Yep the card should perform a little better than the old 550Ti, but it required a new PSU too.
Roughly 400 EUR all in all, but I was tired of trying to fine tune something here and there, now these worries should be gone once the 970 is installed.
Nevertheless, if only out of curiosity, I'll try to see whether I can squeeze out a little more from the 550Ti.

Best regards - Mike

[rruff]

PSU replacement is more involved, but you might be fine with your current PSU. TDP of the 970 is only 30W more than the 550Ti. Impressive efficiency. Slower AMD cards are using 2x the power.

[rruff]

If anyone here is shopping this is the best deal I've seen on a GTX 750... provided that you do the rebate ($30) and sell the game currency on ebay (been $15+ lately) you'll end up ~$50 out of pocket. I hate rebates too, but...

http://slickdeals.net/f/7435552-gtx-750-65-ar-app-fs-tiger-direct-pny-xlr8-geforce-plus-in-game-currency-selling-for-15-on-ebay-50-out-of-pocket?v=1

EDIT: The rebate is $50 today! (Fri Nov 28). Practically free now. Hard to believe they can pay all those rebates, since they are surely losing money when they do.

[SAS~Storebror]

That indeed sounds like a good deal.

Last night I did some more testing and good interesting results. Interesting because it wasn't quite what I expected to see. Turns out that FXAA isn't the best performer and indeed the game applies 4x anisotropic filtering when you set conf.ini accordingly, even more, any higher AF setting in the Nvidia Control panel doesn't seem to change much.

Regarding "application controlled anisotropic filtering", this was my initial issue where I've had ugly graphics with that setting, see the missing details on the front of the carrier deck:


However this was just a wrong setting in Nvidia Inspector and/or the conf.ini, after resetting everything to defaults (inspector) and "perfect mode" settings (conf.ini) I got this for ingame anisotropic filtering:


That looks exactly like 4xAF, and funny enough, even 16xAF doesn't change this (the additional smoothness in this image comes from other AA settings):


Now to the results from video track testing ("The Black Death"), all tests have been performed on the same system running the first 90 seconds of that thread, game version was Ultrapack 3.0 RC4 with no mods.

Settings	FPS AVG	FPS MAX	FPS MIN	FPS MIN (*)	Remarks
FXAA (ingame AF)	117	161	55	31	1st run of the game
./.	118	160	89	45	2nd run of the game
./.+TB+AFO+ASO	118	161	89	45
./.+4xAF	118	163	90	46
./., ingame AF disabled	76	146	37	37	looks horrible
8xCSAA+TB+AFO+ASO+ingame AF	121	203	81	45	better than FXAA!!
4xMSAA+./.	127	209	88	45	best AA match for 550Ti
./.+1xSSTAA	127	212	87	45
./.+2xSSTAA	88	134	62	42	too much for 550Ti
./.+4xSSTAA	55	90	37	37	way too much for 550Ti
./.+2xSpGSSTAA	127	209	88	45
./.+4xSpGSSTAA	127	209	88	46
./.+8xSpGSSTAA	127	210	88	45	best TAA match for 550Ti
./. (inkl.TAA) +16xAF	127	211	88	45	No AF impact on FPS ?!?

Red = Setting dropped during testing for either performance or quality reasons.
Green = Setting kept for further tests.
AVG = Average
FPS MAX = Maximum FPS during test run
FPS MIN = Minimum FPS during test run, expect situation below
FPS MIN (*) = FPS impact when incendiary bomb is dropped on care column at ~70 seconds of the video.
FXAA = Nvidia Full Screen Anti Aliasing.
AF = Anisotropic Filtering
TB = Triple Buffering
AFO = Anisotropic Filter Optimization setting in Nvidia Inspector
ASO = Anisotropic Sample Optimization setting in Nvidia Inspector
CSAA = Coverage Sample Anti Aliasing
MSAA = Multisample Anti Aliasing
TAA = Transparent Anti Aliasing
SSTAA = Supersampling TAA
SpGSSTAA = Sparse Grid Supersampling TAA

Some preliminary conclusions (needs more testing to verify):

• FXAA looses against 4xMSAA both in terms of quality and performance. 8xCSAA is nearly on the same performance level, quality needs further testing.
• normal SSTAA is a no-go for the 550Ti.
• SpGSSTAA shows no performance penalty at any setting. Impressive. Quality needs further testing.
• AF settings in Nvidia Control panel seem to have no effect ingame (??!?).
• When the game runs for the first time, FPS is reduced for the need of loading stuff off the HDD.
• When the incendiary bomb is dropped on the car column, the explosion effects seem to hit some other limit than the GPU, it's "virtually" nailed at max. 45 FPS regardless FPU settings.

Best regards - Mike

[Mick]

... and indeed the game applies 4x anisotropic filtering when you set conf.ini accordingly ...

Mike, could you please tell us what  these settings are in the conf.ini ...? 

[SAS~Storebror]

Oh yes, good catch.
The responsible lines in conf.ini are:

Code: [Select]

[Render_OpenGL]
TexMipFilter=3
TexFlags.TexAnisotropicExt=1
This gives you ingame anisotropic filtering.

Best regards - Mike