GTX 680 and Crysis 2

[computeguy]

So, I have a fairly new GTX 680 2gb, and I wanted to play Crysis 2 on it. I read on many benchmarks that it should be getting 70-75 average FPS on ultra DX11. I am only getting around 50 FPS average with those same settings. I feel like I have a faulty card. My CPU is definitely not a bottleneck, as it's an i5 2500k. I do have the latest drivers installed (314.07 WHQL). Also, is Crysis 2 supposed to be taking up a lot of VRAM? Up to 1.97gb? I find this so hard to believe because I see people running Crysis 3 at Ultra with SMAAx4 on a 2gb card. What do you guys think? Is it faulty?

 

[unksol]

Those benchmarks are generally done on overclocked sandybridge E processors. A 2600K at stock can certainly bottleneck a 680 in some games.

What is your monitor resolution, PSU, system memory etc? What are your temps and CPU/GPU usage? If your GPU usage.is not near 100% it's bottlenecked or throttling.

 

[helper800]

Also most benchmarks are done in single player and I have a gut feeling that you are online testing the fps. If you are testing online then the next thing to do would be to test in single player and compare.

 

[computeguy]

Those benchmarks are generally done on overclocked sandybridge E processors. A 2600K at stock can certainly bottleneck a 680 in some games.

What is your monitor resolution, PSU, system memory etc? What are your temps and CPU/GPU usage? If your GPU usage.is not near 100% it's bottlenecked or throttling.

Sorry, late response. I'm at 1080p, with a HX850 (for future SLI), and 8gb of 1600mhz Corsair VENGEANCE Ram. CPU temps are fairly low, around 38-40C. GPU temps are also not bad, usually at 72C. I'm not sure of my GPU usage, I'll have to check that.

 

[computeguy]

Also most benchmarks are done in single player and I have a gut feeling that you are online testing the fps. If you are testing online then the next thing to do would be to test in single player and compare.

I am testing in single player.

 

[helper800]

Definitely post back with gpu usage and usages on each individual core on your cpu. If you get less than 90% on your gpu usage and 100% on any one core on your cpu you are getting a cpu bottleneck.

 

[computeguy]

Definitely post back with gpu usage and usages on each individual core on your cpu. If you get less than 90% on your gpu usage and 100% on any one core on your cpu you are getting a cpu bottleneck.

GPU usage seems perfectly normal. It's 99% when under 60FPS (it's 60 becuase I have it locked there) and obviously it goes down once it's at 60+, like when I'm staring at a wall (because it doesn't need to use all of the power, right?)

 

[helper800]

GPU usage seems perfectly normal. It's 99% when under 60FPS (it's 60 becuase I have it locked there) and obviously it goes down once it's at 60+, like when I'm staring at a wall (because it doesn't need to use all of the power, right?)

Test without v-sync on. Then your average FPS will more than likely be on par with what your cpu is as compared to the bench marking peoples' sandy-bridge E processors.

 

[computeguy]

Test without v-sync on. Then your average FPS will more than likely be on par with what your cpu is as compared to the bench marking peoples' sandy-bridge E processors.

V-sync was off :/ I just had it locked to 60 using EVGA Precision. I also played it a bit with it set to unlimited FPS and no Vsync, and was still getting the same results.

 

[helper800]

V-sync was off :/ I just had it locked to 60 using EVGA Precision. I also played it a bit with it set to unlimited FPS and no Vsync, and was still getting the same results.

This is odd. I have no further knowledge on why it would be so much lower than the benchmarks. But personally i cannot tell the difference between 60 and 50 fps and i'm sure if you could the difference is minute. As for the fps drops because you are not getting performance on par to the benchmarks you cannot do anything about it sorry. Only explanation I can think of would be that they are using sandy bridge E processor's that are obviously highly overclocked to ensure no bottle neck of any kind exists during benchmarks to ensure the results reflect the true power of the GPU, game settings and resolution at hand.

 

[dudewitbow]

also note that performance can jump around as boost clocks inflates and can vary scores from one site to another. You also have to double check the exact settings of the benchmark.

 

[dudewitbow]

also note that performance can jump around as boost clocks inflates and can vary scores from one site to another. You also have to double check the exact settings of the benchmark.

 

[redeemer]

I think 70-75 on a single GTX 680 in Crysis 2 ultra is too generous! Of course this depends on your resolution setting??

About Crysis 3 sure you can run 4X SMAA Ultra setting on a single 680 however the game will not be smooth and barely playable.

 

[computeguy]

This is odd. I have no further knowledge on why it would be so much lower than the benchmarks. But personally i cannot tell the difference between 60 and 50 fps and i'm sure if you could the difference is minute. As for the fps drops because you are not getting performance on par to the benchmarks you cannot do anything about it sorry. Only explanation I can think of would be that they are using sandy bridge E processor's that are obviously highly overclocked to ensure no bottle neck of any kind exists during benchmarks to ensure the results reflect the true power of the GPU, game settings and resolution at hand.

There is the possiblity of it being slightly defective, right? I would like my GPU to preform as it should, that's why I'm not happy with 50 average. Also, I think I read somewhere that ~4.5Ghz (probably around what the benchmarking people were at, if they did OC) will barely increase the performance of games like Crysis, as they rely a lot more on the GPU.

 

[GameLifter]

One thing to note is that most benchmark sites use a benchmark tool to benchmark Crysis 2. When I ran it on mine with just one GTX 680 my score was is in the lower 70s but during actual gameplay I would get severe fps drops in areas with lots of particle effects. I did some research and found this was common due to the crazy particle system in Crysis 2. Even adding a second GTX 680 for SLI didn't fix the fps drops. What's ironic is that Crysis 3 runs better with my system overall with two GTX 680s than Crysis 2 thanks to better scaling and maybe better driver and game optimizations. Hope this helps!

 

[computeguy]

I think 70-75 on a single GTX 680 in Crysis 2 ultra is too generous! Of course this depends on your resolution setting??

About Crysis 3 sure you can run 4X SMAA Ultra setting on a single 680 however the game will not be smooth and barely playable.

I'm at 1080p.
Crysis 3 Vid:
http://www.youtube.com/watch?v=xDOdBJ5LxUA&list=UUftcLVz-jtPXoH3cWUUDwYw&index=1
I'm sorry, I meant High 2x SMAA on a 680...yeah, big difference.

 

[computeguy]

I think 70-75 on a single GTX 680 in Crysis 2 ultra is too generous! Of course this depends on your resolution setting??

About Crysis 3 sure you can run 4X SMAA Ultra setting on a single 680 however the game will not be smooth and barely playable.

Also, here is another video of actual gameplay of Crysis 2 on a 680. Description says 60+ in all areas, Ultra, high res textures, DX11 upgrade.
http://www.youtube.com/watch?v=KxN5_96M2TE