Cryostasis Gameplay Performance and IQ @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,635
Cryostasis Gameplay Performance and IQ - Cryostasis arrived from Russia this spring, mostly to hushed anonymity. NVIDIA seems to want it to be a PhysX showcase, but is it really up to the job? Come with us as we find out, using seven of today's best video cards.


Ultimately, however, Cryostasis is just not a good game with which to judge a video card upgrade. It is worth playing, but don't buy a new video card for it.
 
Nice article. I am really looking forward to seeing how much of an effect an additional GPU will help with physics. Like pairing my GTX260 with my 8800GTS for physics. Can't wait. Thanks!:cool:
 
The game reminded me of bioshock, except with better lighting and effects.
 
Absolutely atrocious graphic to video power ratio. From the sc's there is no justification at all why you would need a top end card to run this thing at even medium set ups. You would think after the whole Crysis incident the devs would have learned a thing or two.
 
Great, I just bought a 2nd GTX 280 to SLI, and now I can only play lowest settings?

This is crazy. I might have to lower my resolution for gaming :-( No. No way. 2560x1600 or death.

I wonder if GTX 280 SLI will be better than GTX 295?

Anyone with insight on that?
 
Absolutely atrocious graphic to video power ratio. From the sc's there is no justification at all why you would need a top end card to run this thing at even medium set ups. You would think after the whole Crysis incident the devs would have learned a thing or two.

Except that Crysis actually gives you a visual payoff for your investment.

Cryostasis does not.

Ultimately, this is a linear, single player, story-driven game. Its replayability will thus be limited to those who A) like the story, and B) like reading the same book over and over....and C) those who have pretty high-end gear.
 
I hear that it is a very scary and atmospheric game for the first time through, but once the novelty and original premise wear off it becomes a bit by-the-numbers (doesn't help that supposedly the shooting itself sucks). In addition to the performance issues, the game also has absolutely atrocious widescreen support. Doesn't look like a bad buy, especially if it drops in price (I know some people really like the game), but definitely not what I'd call a must-have.
 
I hear that it is a very scary and atmospheric game for the first time through, but once the novelty and original premise wear off it becomes a bit by-the-numbers (doesn't help that supposedly the shooting itself sucks). In addition to the performance issues, the game also has absolutely atrocious widescreen support. Doesn't look like a bad buy, especially if it drops in price (I know some people really like the game), but definitely not what I'd call a must-have.

Pretty much. The shooting is somewhere between accurate and S.T.A.L.K.E.R.

But the first time you encounter a zombie, you don't have a weapon, so it is pretty...intense.
 
Except that Crysis actually gives you a visual payoff for your investment.

Cryostasis does not.

Ultimately, this is a linear, single player, story-driven game. Its replayability will thus be limited to those who A) like the story, and B) like reading the same book over and over....and C) those who have pretty high-end gear.

It's nice to see someone actually mention this, theres several discussions in the games forum about performance and optimization. I've seen the visuals for this game and it isn't doing anything special, my guess would be very bad occlusion - to test that you could enable wireframe in the game and see whats being rendered, it could be that a decent portion of the level is being drawn even if you can't see it yet. Custom engines from small studios tend to lack in these kind of departments.

I agree with Crysis, it's demanding but it's got very high visual fidelity, so theres payoff for the performance kick to the nuts. People whine about poor optimization and this game is a perfect example, Crysis isn't at least no one has really provided any proof that it is, we can find plenty of games like this that look better and run better, like bioshock.
 
Error Fixed - Thanks, Brent
 
Last edited by a moderator:
I noticed the same thing you guys did about performance vs. graphical detail when I played the demo. I thought the graphics looked like something that would have come out three or four years ago. I'd rather play Doom 3. It'll make you jump and the graphics and performance are better to boot.
 
I still have a BFG pci Physx card laying around. I would be curious if it would improve the ATI results if you used the original Physx add-on card.
 
I still have a BFG pci Physx card laying around. I would be curious if it would improve the ATI results if you used the original Physx add-on card.

i would also like to know this. how would ati perform on an even playing field?
 
Absolutely atrocious graphic to video power ratio. From the sc's there is no justification at all why you would need a top end card to run this thing at even medium set ups. You would think after the whole Crysis incident the devs would have learned a thing or two.

One thing about Crysis, that was different in this game, is that with Crysis DX10 was slower, but in Cryostasis DX10 is faster, so at least it got that part right.
 
I still have a BFG pci Physx card laying around. I would be curious if it would improve the ATI results if you used the original Physx add-on card.

i would also like to know this. how would ati perform on an even playing field?

It could improve the situation a bit, but I don't have a great deal of confidence that it would make a big difference. I just don't think the game is built very well.
 
The game reminded me of bioshock

Don't know how you come up with that at all. Not even close. Bioshock kicked the crap out of this one in every way.

I just did not like this game one bit, I thought it was just plain terrible. There wasn't one single thing about it that I liked. Just not my cup of Russian tea.
 
I'm currently playing Cryostasis with everything maxed at 1280x1024 and is completely playable, for some odd reason, turning on or off PhysX doesn't make a difference in performance and some stuttering is easy to spot when it is running in DX10, the same issue that I first saw in Far Cry 2, to eradicate that issue, I just play the game, minimize it, open another 3D app like Ruby's DX10 demo, run it for 5 seconds, close it, maximize the game and presto, quite odd.
 
I'm currently playing Cryostasis with everything maxed at 1280x1024 and is completely playable, for some odd reason, turning on or off PhysX doesn't make a difference in performance and some stuttering is easy to spot when it is running in DX10, the same issue that I first saw in Far Cry 2, to eradicate that issue, I just play the game, minimize it, open another 3D app like Ruby's DX10 demo, run it for 5 seconds, close it, maximize the game and presto, quite odd.

I would guess it's because your resolution is so low. The PhysX won't matter because you don't have a physx card.

Next is the Hardware physics option. Toggling this option enables NVIDIA’s PhysX technology to be rendered on any PhysX capable device in the system, whether it be a GeForce 8+ series GPU or an AGEIA PhysX standalone PCI card. If you do not have a PhysX capable device in your PC, this option does nothing.


That's a good write up. I was considering buying it just for the storyline/suspense, but I doubt I will now. Thanks for saving me 60 bones.
 
i would also like to know this. how would ati perform on an even playing field?

That wouldn't be an even playing field, because the GeForce would be doing both graphics and physics, while the Radeon would only be processing graphics, with the PPU working on physics. The only way to be even, would be to have a GeForce for graphics only and a PPU for physics,
However that's not the point, since we want to see how the GPU alone can handle both tasks. Given that only GeForces can process PhysX and this is a PhysX game, tough luck for Radeons.
 
So the hyped PhysX game is not worth the hype?
Let’s put this simply: Based on what we saw of the level of graphical detail in Cryostasis, we really don’t see any good reason for performance to be so wretched. The interior geometry is plain, to say the least, and the outdoor geometry is not even as detailed as what we saw several years ago in Doom 3. The textures on the creatures are nicely detailed, but the world geometry textures are flat and featureless in the extreme. We are fairly certain that the problem is related to the ever-present creeping frost texture/shader/trick that is applied to virtually every surface and object within the game (to include the player’s weapons). But even that isn’t terribly impressive. It looks nice up close, but from a distance it could be replaced with a simple, flat frosty texture and we doubt anybody would notice the difference. The frost is an interesting feature, but not one that should destroy performance.
 
Last edited:
I find it more interesting, that the system is not even making heavy full use of the CPU, when it should have offloaded some of the physics work since there was so much free horsepower.
 
Love the highest resolution playable at max settings. Very telling chart. Pretty much told me, do NOT buy this game. My 4850 cold probably play it at 800x600...
 
Those screenshots look like ass. Does the game really look that bad? Looks like a totally unoptimized piece of crap. Of course I'm jumping to conclusions here, I haven't seen it live myself.
 
Those screenshots look like ass. Does the game really look that bad? Looks like a totally unoptimized piece of crap. Of course I'm jumping to conclusions here, I haven't seen it live myself.

Um...yes.
 
Damn when are they releasing the patch that fixes the game?

So far it has been a fun time with the game (yet at low fps).
 
Love the highest resolution playable at max settings. Very telling chart. Pretty much told me, do NOT buy this game. My 4850 cold probably play it at 800x600...

Damn why are people having such a hard time running this game?

I am running it at 1440X900 (maxed with no aa, game doesnt seem to need it) with a HD4870 512 and I am getting 25-50fps at all times (32 fps average)
 
I can't believe Nvidia, there STILL isn't a phsyx game that I am willing to buy. I certainly understand that nvidia is doing a better job at vendor relations but if this is what they are coming up with I can see why ATI is not worried. I think Nvidia needs a reality check here, looking at some of the up and coming games (ghostbusters comes to mind) and looking at this game why do they think physix is still a selling point?

Nice review guy, though I am not sure I would be so hard on AMD not being on board, it doesn't look like they are missing much (here at least)
 
I am running it at 1440X900 (maxed with no aa, game doesnt seem to need it) with a HD4870 512 and I am getting 25-50fps at all times (32 fps average)

Resolution matters *alot* in this game, and yours has less pixels than even 1280*1024. Also, in most areas there aren't much heavy physics going on. In the areas there are, you should see a significant fps drop.
 
Great, I just bought a 2nd GTX 280 to SLI, and now I can only play lowest settings?

This is crazy. I might have to lower my resolution for gaming :-( No. No way. 2560x1600 or death.

I wonder if GTX 280 SLI will be better than GTX 295?

Anyone with insight on that?

Perhaps I'm a bit late but...
2 gtx 275's perform slightly better than a single 295, and the 280's perform slightly better than the 275, though not really worth the price difference there. So 2 280's are better than a 295.
 
Resolution matters *alot* in this game, and yours has less pixels than even 1280*1024. Also, in most areas there aren't much heavy physics going on. In the areas there are, you should see a significant fps drop.

I am aware of that, its running bad but not that bad.
 
Nice review guy, though I am not sure I would be so hard on AMD not being on board, it doesn't look like they are missing much (here at least)

True, but it seems to me that AMD/ATI misses the boat on almost every game.
 
True, but it seems to me that AMD/ATI misses the boat on almost every game.

What you are stating here doesn't make sense at all. There are fewer tittles with GPU accelerated PhysX than DX10.1 tittles, so if you have something smart to say please tell me and enlight us all, if don't, then you are just trolling.
 
What you are stating here doesn't make sense at all. There are fewer tittles with GPU accelerated PhysX than DX10.1 tittles, so if you have something smart to say please tell me and enlight us all, if don't, then you are just trolling.

His points are explained in the article that he authored. He is not trolling, but rather forming and sharing an informed opinion. Maybe you do not share that opinion, and if you don't, we would love to hear that.

As for DX10.1, we have done plenty of articles on that as well.
 
What you are stating here doesn't make sense at all. There are fewer tittles with GPU accelerated PhysX than DX10.1 tittles, so if you have something smart to say please tell me and enlight us all, if don't, then you are just trolling.

DX10.1 offers about as much for gameplay as PhysX does, in my opinion. (Little-to-nothing, in case you need it spelled out.)

The point is, if you add up the TWIMTBP titles and compare them to titles with ATI logos on the boxes, you will notice that developer relations are pretty one-sided, regardless of DX10.1 or PhysX support.
 
DX10.1 offers about as much for gameplay as PhysX does, in my opinion. (Little-to-nothing, in case you need it spelled out.)

The point is, if you add up the TWIMTBP titles and compare them to titles with ATI logos on the boxes, you will notice that developer relations are pretty one-sided, regardless of DX10.1 or PhysX support.

Now I understand, it always has been like that even when ATi dominated the GPU market, now that It doesn't, it will be more difficult, ATi should change it's strategy, considering that the current ATi architecture is quite dependant of optimizations.
 
Game is not multi core too.. explains why so crap fps.

I wonder if that is the case as well.
CPU usage often looks like it is spread out over many cores when it is just running one core.


btw, thanks [H] for an amazingly detailed review, it answered everything I wanted to know bar the use of dedicated PhysX cards.
I am looking forward to your update with PhysX cards.

Would it be possible to include the 8800GT in the dedicated PhysX tests?
I (and am sure many others) have a spare one lying around that might be useful here.
Cheers.
 
The performence hit could just be from the normal maps being too large for what they are achiving. An easy test of that would be to export the DDS files to photoshop resize them by half and see what impact it has.
 
Back
Top