Benching the Benchmarks: STALKER – Clear Sky DX10.1 Benchmark

Dx10.1, “MSAA for A-tested Objects” and Conclusion:

Do you think it can get worse?  Oh yes.  More Clear Sky DX10.1 features will make every card struggle now – at all resolutions!

In game, night

In game, day

If you thought DX10 was demanding on your PC on the “regular” DX10 pathway for Clear Sky, wait until we check “MSAA for A-tested objects” – which requires still higher performance from your video card.  Notice that the 280GTX takes a bigger performance hit – percentage-wise – than the 4870.  Also, we note there is an additional highest setting available for Radeon cards that is not available to Nvidia cards – DX10.1.  Fortunately for ATi owners,  when we use a powerful card like 4870×2, it’s performance hit is slightly less than with the lower setting that the GTX280 Geforce can run.  In actually playing the Clear Sky game, however, the practical difference in visuals between DX10.0 and DX10.1 is unnoticeable; especially when you are being shot at.

Here is our final set of four benchmarks – all at 1920×1200 to show the performance variablity between out test cards.  The “+” is MSAA for A-tested objects and Dx10.1 is only run on 4870 and 4870-X2 as the option for DX10.1 is grayed out for Nvidia cards.  We did not bother with and retired our 8800GTX for this test.

Conclusion:

You might have noticed that CrossfireX-3 – a staple of our “GPU Shootout Series”– is conspicously missing.  That is because it refuses to run with either the CS benchmark – or the Clear Sky expansion –  since it was released.  We hope AMD is paying attention as this game needs all the graphics muscle it can muster.  Clearly, this generation of video card is not well suited for playing Clear Sky even at 1920×1200 with fully maxed out DX10 settings.  The only hope is for multi-GPU and well scaling drivers.

Realistically, unless you are playing with a 4870×2, you are going to have to turn down some settings in CS – even at 1680×1050 resolution.  We will keep you updated if drivers improve this situation for any of our test cards – and expecially for CrossfireX-3.  In the meantime, we hope you enjoy running this benchmark for yourself.

Feel free to post comments below or discuss this review in our ABT forums.

Mark Poppin

ABT Editor