Remember Me?
   Lost your password?

Chernobyl Shooter tested with 11 graphics cards

Stalker Clear Sky: DX10 Benchmark Review

Since Friday Stalker: Clear Sky in on sale. But with the unpachted version it wasn't possible to test the game appropriately. The update to version 1.5.03 stabilized the game and removed enough bugs to make benchmarks possible. The result is a comparison between eleven graphics cards and three different resolutions.

Visual differences
Stalker: Clear Sky is based on the X-Ray Engine of the predecessor Shadow of Chernobyl. Among enhanced visual effects and better textures, the new Direct X 10 support is advertized. The differences between DX9 and DX10 can be seen in our exclusive screenshot comparison.

But in the game the new effects are visible in certain situations only. Due to the enhanced lighting of objects in Direct X 10, Clear Sky looks more realistic but compared to DX9 the performance doesn't differ very much in our benchmarks. The only option to deliver some noticeable visual differences is the wet surfaces - but they don't cause a real loss of performance either.

Stalker Clear Sky: DX10 Benchmark Review
[Source: view picture gallery]

Therefore the two versions (DX9 and DX10) are identical in regards of the required hardware but not in regards of the visualization.

(Video) Memory requirements
We decided to test all cards on Direct X 10 and Windows Vista x64, in order to receive the full benefit of 4 GiByte RAM - and Stalker: Clear Sky likes it. The game uses up to 1.6 GiByte memory after some time. In combination with the memory required by the operating system, this exceeds the capabilities of systems with only 2 GiByte RAM and therefore lags are inevitable.

This problem gets worse because of the game's intense usage of video memory. Up to 800 MiByte in DX9 were not uncommon and, according to the console, in DX 10 even more VRAM was used. If you have a graphics card with only 512 MiByte memory, the remaining 300 MiByte are transferred to the system memory. Because of this we recommend 3 GiByte (Windows XP) or 4 GiByte (Windows Vista).

Processor requirements
Stalker: Clear Sky doesn't use more than a single core. Solely the graphics driver sometimes uses an additional CPU core sometimes. Therefore gigahertz and the performance per clock are important.

Stalker Clear Sky: DX10 Benchmark Review
[Source: view picture gallery]

Author: Marc Sauter (Sep 08, 2008)


Comments (15)

Comments 12 to 15  Read all comments here!
cloud_strife1 Re: Stalker Clear Sky: DX10 Benchmark Review
Junior Member
27.04.2009 06:19
The only problem I notice is that when I go to load a saved game it takes FOREVER to load, maybe about 4 minutes, but it's fast before and after that.

I'm running Windows 7 64-Bit Ultimate, AMD Athlon 64 X2 5000+, 250GB Maxtor SATA 300GB/S HDD. 2 GB of DDR2 with a HIS 512MB GDDR3 ATI Radeon HD 4670 card and my framerates never drop below 35ish even during GPU intensive moments. I got the OC version though which is at 780Mhz/2Ghz.
Yapa Re: Stalker Clear Sky: DX10 Benchmark Review
Senior Member
16.09.2008 05:51
The slowdown during morning and evenings could be due to a engine bug/loop rather than not enough GPU power. I have done a few tests... here is a screenshot which shows the fps drop suddenly at around 6.08am, without any change to the graphics/view. Same results can be seen when just staring at a wall... fps drops even tho the GPU is not under load

more info:
senseless Re: Stalker Clear Sky: DX10 Benchmark Review
Junior Member
14.09.2008 14:01
Thanks a lot for review!

As you can see we have 100% load on 1st CPU core and 0-10% on 2nd and so on with multicores CPU used. Just want add some hint about this trouble:

- start game and wait for intro start playing or game menu displayed
- press CTRL+SHIFT+ESC for call Task Manager (if it doesn't try ALT+TAB, WINKEY or CTRL+ALT+DEL and start it way you do)
- in Task Manager go to "processes tab" and find xrEngine.exe press RMB on it then select "Affinity" - you will see that all cores are used
- unselect all cores except 1st press "OK"
- now enter xrEngine.exe Affinity settings again and select all cores back to use then press "OK"
- press ALT+TAB to switch back in STALKER process
- now play and see the difference on ingame cores load - on AMD/Intel multicore CPUs i have seen 50-70% for each core (you may use RivaTuner OSD or switch back to TaskManager to see CPU load history)

And another one hint:
- on ATI cards is great performance boost measured when using Catalyst not 8.8final but 8.8beta.
- so it able that x86 vs x64 systems will give performance differense
so try yourself

Copyright © 2015 by Computec Media GmbH      About/Imprint  •  Terms/Conditions