Graphic benchmark results

Graphic benchmark results

Postby Calinou » 28 Jul 2014, 11:49

Hardware:
Processor: Intel Core i7-2600K @ 3.40GHz (4 cores, 8 threads)
Memory: 8192MB, Disk: 256GB Samsung SSD 840
Graphics: MSI NVIDIA GeForce GTX 570 1280MB (732/1900MHz)

Software:
OS: Ubuntu 14.04
Kernel: 3.13.0-32-generic (x86_64)
Desktop: Xfce 4.10
Display server: X Server 1.15.1
Display driver: NVIDIA 331.38
OpenGL: 4.4.0
Screen resolution: 1920 × 1080


Game is in 1920 × 1080, fullscreen. Compositing (desktop effects) of Xfwm is enabled. The track used is I6-Flooded, at start position with default camera settings.


Default settings (the ones used when you first start the game):

110 FPS. 117 FPS with static car reflections (+ 7). Disabling water reflection and reflection goes to 140 FPS (+ 23). Disabling shadows goes to 178 FPS (+ 38).
FPS with different shadow sizes, from lowest to highest: 146 FPS, 145 FPS, 142 FPS, 138 FPS, 122 FPS (with shadow count 2). Thus, 1024 shadow size is the most efficient one. Using 1 shadow instead of 2 makes distant shadows uglier, but is only slightly faster. Using 3 shadows is slower for no benefit.


FPS with various graphic presets (after a track restart, so all presets used the high-quality textures):

Lowest: 431 FPS
Low: 371 FPS
Medium: 194 FPS
High: 135 FPS
Higher: 100 FPS
Very high: 75 FPS, without effects: 99 FPS
Ultra: 45 FPS, without effects: 66 FPS
Impossible: 25 FPS, without effects: 53 FPS

Playing the game with Impossible preset is actually possible, Stunt Rally doesn't need very high FPS after all.

Effects:

Lowest + all effects: 42 FPS
Lowest + all effects but god rays: 80 FPS

“God rays” is a very expensive effect (disabling it gives you + 38 FPS), it's buggy anyway and generally does nothing.

Vegetation:

It seems to be hard to get reliable data on the performance impact of vegetation, but using the “Impostors only” setting will be quite faster, almost as fast as disabling all trees. However, with this setting, trees won't cast any shadows.


You too can try contributing by making your own benchmark like that one. Just test all presets, changing from one preset to another, then pressing F5 to reload the track.
Last edited by Calinou on 25 Jan 2015, 12:08, edited 1 time in total.
User avatar
Calinou
 
Posts: 161
Joined: 22 Jan 2010, 21:43
Location: France

Re: Graphic benchmark results

Postby CryHam » 28 Jul 2014, 14:03

This is good. And informative.
But for true comparison of modes I recommend quitting after each preset change (it is mandatory, not obeying may result in damage and other loss, for which I can't be made responsible LOL :lol:). E.g. texture size doesn't change, needs full quit, etc.

And I recommend some complicated track like F12,J13, I6 is quite empty and really high Fps all the time.
Anyways keep in mind that Fps depends on everything now, how many cars (ghosts) are visible too, etc.
Maybe only except weather outside, but I'm nearly sure solar flares do affect SR Fps.
And lastly, this is not a hardware commercial and neither OS.

Anyways, my results, we have similar hardware.

Hardware:
Processor: Intel Core i7-2600K @ 4.40GHz (4 cores, 8 threads)
Memory: 8192MB, Disk: 128GB OCZ Agility3 SSD
Graphics: MSI GeForce GTX 560 Ti 1024MB (880/1760MHz)

Software:
OS: Windows 7 64-bit
Display driver: NVIDIA 335.23
Screen resolution: 1920 × 1080 windowed


Both ghosts disabled. Car: ES, tracks not reversed. Antialiasing: 0.

On start place of I6-Flooded:
game default (higher, no effects) settings
213 FPS.

Lowest: 825 FPS, but black terrain IDK
Low: 720 FPS, but black terrain far
Medium: 446 FPS
High: 276 FPS
Higher: 191 FPS
Very high: 130 FPS, without effects: 194 FPS
Ultra: 60 FPS, without effects: 101 FPS
Impossible: 32 FPS, without effects: 72 FPS


On the start place of F12-HighPeaks.

Lowest: 722 FPS, black terrain
Low: 484 FPS, but black terrain far
Medium: 273 FPS
High: 162 FPS
Higher: 63 FPS, without effects: 67 FPS
Very high: 48 FPS, without effects: 59 FPS
Ultra: 26 FPS, without effects: 34 FPS
Impossible: 15 FPS, without effects: 21 FPS
User avatar
CryHam
SR Moderator
 
Posts: 695
Joined: 25 Nov 2012, 08:40

Re: Graphic benchmark results

Postby CryHam » 01 Oct 2014, 17:04

Ugh. After talking with Calinou it seems that he is using global anisotropy at value 16. At game's default Higher it is just 4.
This itself could be already 2x less Fps, so yeah I'm gonna say his results are completely invalid, and my Kubuntu install was gone lately, sorry.
User avatar
CryHam
SR Moderator
 
Posts: 695
Joined: 25 Nov 2012, 08:40

Re: Graphic benchmark results

Postby CryHam » 24 Jan 2015, 19:33

I got my Kubuntu back so, here are my results from ver 2.5 binary.
I didn't have sound though, from IDK what reason.

Hardware:
Processor: Intel Core i7-2600K @ 4.40GHz (4 cores, 8 threads)
Memory: 8192MB
Graphics: MSI GeForce GTX 560 Ti 1024MB (880/1760MHz)

Software:
OS: Kubuntu 14.04
Kernel: 3.13.0-44-generic (x86_64)
Desktop: KDE 4.13.3
Display Server: X Server 1.15.1
Display driver: NVIDIA 331.113
Screen resolution: 1920 × 1200 windowed


Both ghosts disabled. Car: ES, tracks not reversed. Antialiasing: 0.

On start place of I6-Flooded:
game default (higher, no effects) settings
138 FPS.

Lowest: 573 FPS
Low: 472 FPS
Medium: 276 FPS
High: 181 FPS
Higher: 122 FPS
Very high: 87 FPS
Ultra: 42 FPS
Impossible: 22 FPS


On the start place of F12-HighPeaks.

Lowest: 514 FPS
Low: 385 FPS
Medium: 188 FPS
High: 112 FPS
Higher: 59 FPS, without effects: 63 FPS
Very high: 44 FPS
Ultra: 22 FPS
Impossible: 13 FPS
User avatar
CryHam
SR Moderator
 
Posts: 695
Joined: 25 Nov 2012, 08:40

Re: Graphic benchmark results

Postby Calinou » 26 Jan 2015, 18:28

dimproject {l Wrote}:Kubuntu 14.04 a little old and not actual.


Performance should nearly be the same in comparison to 14.10.
User avatar
Calinou
 
Posts: 161
Joined: 22 Jan 2010, 21:43
Location: France

Who is online

Users browsing this forum: No registered users and 1 guest