Has anyone seen this?
http://www.pcgameshardware.com/aid,7...tips/Practice/
Has anyone seen this?
http://www.pcgameshardware.com/aid,7...tips/Practice/
It has been out for a while...a useful article for a general idea about what is going on, but since then both the game patches and vid driver updates have changed the performance landscape, at least in the testing I've done.
The other thing that is missing in the article is the difference between fps for soldiers daydreaming and the heat of battle with smoke and explosions in the middle of a city or fort. Not only are those situations a drain on performance, but they change the relative nature of all those gfx settings. As well, the hight and distance of the camera can make a big difference...for example, any AA setting above 0 heavily taxes close up shots within smoke, but doesn't do hurt fps much with an arial view. Therefore, how you play the game, how you watch it, or how you record it will paint a wholly different fps picture.
In sum, interesting article for understanding some settings basics...but getting a pen and paper and FRAPs is a better start for understanding how your gfx settings impact your gaming experience.
cheers!
It maybe as dated as CA's annonuced specs but it is better than I have seen elsewhere.
If anyone has run tests and willing to share the results, I believe it would benefit all.
very true...I like their pics where you click and see the difference...very helpful
My tests are all on post-it notes LOL and specific to my system. I should organize them and post them I suppose, in case someone has a rig like mine. Maybe if I get my wife some flowers, she'll organize it for me![]()
I see this long time ago, here is also for Empire Total War
http://www.pcgameshardware.com/aid,6...CPUs/Practice/
Test results for FPS penalties and settings
Rig for test:
i7 930 2.8 quad (OC to 3.2)
Windows7 64-bit Pro
6 gig DDR3 OCZ gold RAM12800 1.65v (3x2)
Sapphire 5970 2gig vid card (OC with MSI Afterburner)
ASUS P6T mb
Samsung 23" SyncMaster BX2235
Hitachi HDD 7200rpm 360gb
1 tb iomega external HD
2 tb WD external HD
Game Booster2 Beta3
recording: FRAPS freeware
For testing conditions, I am only using recording the fps while FRAPS is live recording (FPS is red on your screen). This is somewhat different than just standard gameplay, as overall FRAPS costs beaucoup fps. The hits are proportional, therefore a -1 fps recording hit may be higher when not recording.
These numbers are in relation to a "Medium" recording state, starting with the following settings:
RESOLUTION: 1776x1000
SHADER MODEL: high
ANTI-ALIASING: none
TEXTURE FILTERING: trilinear
TEXTURE QUALITY: medium
PARTICLE EFFECTS: medium
UNIT DETAIL: medium
BUILDING DETAIL: medium
SHADOWS: medium
UNIT SIZE: ultra
TREES: medium
GRASS: medium
WATER: medium
SKY: medium
HDR: OFF
VOLUMETRIC EFFECTS: OFF
SSAO: OFF
V-SYNCH: OFF
HARDWARE SHADOWS: OFF
DISTORTION EFFECTS: OFF
DEPTH OF FIELD:ON
Alright, given that medium state, here are the FPS penalties that various changes will bring. Note that most penalties are independent of each other (same penalty no matter what combo), but not all. Anti-aliasing and particle effects are the exceptions, bringing about vastly different, multiplying penalties when used with Volumetric and SSAO. The combinations are many, and not explored here.
These are average FPS penalties while recording the exact unit setup, same map, same starting formation, exact camera angles, with 1/2 stack of ultra unit size in each army.
These tests were run while having several mods active at the same time:
Spoiler Alert, click show to read:
FPS PENALTIES (with FRAPS recording):
RESOLUTION: 1920x1080 -5
SHADER MODEL: high
ANTI-ALIASING: x4 -5
TEXTURE FILTERING: x16 -1
TEXTURE QUALITY: ultra -1
PARTICLE EFFECTS: high -2 **
UNIT DETAIL: ultra -4
BUILDING DETAIL: ultra -1
SHADOWS: ultra 0
UNIT SIZE: ultra
TREES: ultra -1
GRASS: ultra -1
WATER: ultra -1
SKY: ultra -1
HDR+VOLUM+SSAO: ON -5
V-SYNCH: OFF
HARDWARE SHADOWS: ON -1
DISTORTION EFFECTS: OFF
DEPTH OF FIELD:ON
** with occasional stuttering, particularly when N-zoomed
You will see certain settings stayed constant, such as ultra unit size, DoF, v-sync etc. These were not tested because for my recording purposes, they are necessary constants.
With a different computer, of course, your results could vary.
Also note: these penalties, while not exact, apply also to ETW.
Cheers!
Nanny
Last edited by Nanny de Bodemloze; December 12, 2010 at 01:27 PM.
Good informaion, Nanny. Thank you.