also @ TechSpot: California man finds limits of Verizon FiOS unlimited data broadband service: 77TB

Crysis 2 GPU & CPU Performance Test

Discussion in 'Articles and Reviews Comments' started by Julio Franco, Mar 29, 2011.

Post New Reply
  1. slh28 TechSpot Paladin Posts: 1,683   +105

    Would be interesting to see a clock for clock comparison though.
  2. Julio Franco TechSpot Editor Posts: 6,055   +121

    Do you want us to guess? If you didn't read between lines, our take is that they lacked the focus on the PC version development and that's why they didn't go with DX11 or 64-bit executable at launch. Time is always a constraint in software development, but again it's no excuse. Crytek's focus was on launching a multiplatform game and the special PC features were relegated for later.

    Depending on when it happens we probably will. E.g. if it's a few weeks from now, sure, if it's 4 or 6 months later, we may not even bother as a sign of protest.
  3. Jurassic4096 Banned Posts: 158

  4. Per Hansson TS Server Guru Posts: 1,796   +66

    Nice article Steven, enjoyed reading it...
    I just did a little Crysis 2 overclocking test with my Xeon X3370 (=Core 2 Quad 12MB L2 cache CPU)

    3.0Ghz: 45.85fps
    3.6Ghz: 45.92fps

    I think I may be a little bit bound by my old trusty 8800GTS 512MB VGA card even when clocked to 850/1080 (core/mem) :D

    As for my thoughts on Crysis 2 itself, when i played the PC demo I was massively disappointed
    The multiplayer maps above all where way to small, they did not have any of the original Crysis feel to them IMO.
    Now I have not even played multiplayer on my Retail copy of Crysis 2, but after adjusting the Field of View and playing a few levels of the single player mission I must say that I am impressed.
    The visuals are impressive, the lightning and shadows cast, even at the "middle" graphic option of the only 3 available looks amazing.
    It was only now when I read Stevens article that I realized how much better the highest graphical options looks though, however my old trusty VGA card is unable to get any playable framerates there unfortunately.

    However with that said I must highlight that I think Crysis really feels bland without the great Vistas of the original Crysis, there are no such grand views to be seen in Crysis 2 as far as I have played yet, I am now at the mission right after "FDR".

    And how the PC version could be released with options like aim assistance enabled but only 3 preset graphical options with no customization abilities is beyond me.
    And the fact that the console was locked out in the Retail version before the "day one" v1.1 patch just also goes to show where the priorities where during development.
    My biggest gripe is with the default FoV though, I only have a 4:3 monitor but I get headache and feel Nausea only after minutes of gameplay, the Crysis 2 FoV is vertical and the default is 55.
    Settings that work for me and look good with my resolution of 1600x1200 is FoV of 74, with this I get a slight performance reduction but it's well worth it, no longer does it feel like I'm running around looking at the game through binoculars!
    You can use this calculator to figure our your own FoV; http://www.rjdown.co.uk/projects/bfbc2/fovcalculator.php

    The settings to change are;
    cl_fov; this controls the world FoV, you can change it ingame from the console if you have patch v1.1
    pl_movement.power_sprint_targetFov; This controls the sprint FoV, it should match cl_fov otherwise it will appear as you are "zooming in" when sprinting (Since you will be going from a wide FoV to a narrow one)
    r_DrawNearFoV; This controls the FoV of the weapons you use, it is bugged though, if you set it higher than 60 it will go back to 60 when you use the binoculars in the game.

    These are the settings I use in my Crysis 2\system.cfg file, the first two skips the intro movies and multiplayer login screen respectively
    Code:
    g_skipIntro = 1
    g_enableInitialLoginScreen = 0
    cl_fov = 74
    r_DrawNearFoV = 60
    pl_movement.power_sprint_targetFov = 74
    
    Also if you want to run the same benchmark as I did to get the fps numbers above you can download the batch file attached to this post.
    If you are viewing this from the news page comments you need to click on the text "subscribe to this thread" (not tick the button) to get to the forum view...
    Please note that it requires patch version 1.0 (I.e. the Retail release)
    It does not work with v1.1 because Crytek decided to remove almost all CL variables, again a testament to their support for PC gaming, this even made their own included benchmark stop working!

    You can get around this by restoring the backup version of Crysis2.exe under the folder "Backup" in the Crysis2 dir to the bin32 folder, please of course first move the v1.1 Crysis2.exe file out of the way and somewhere you can easily identify it as the 1.1 version...

    Attached Files:

  5. Jurassic4096 Banned Posts: 158

  6. compu4 Newcomer, in training Posts: 46

    I think that Crytek really dropped the ball on this one from a technological standpoint. Here is my list of grievances:


    1. The game only has DX9 support. That is a serious error on Crytek's part. The original Crysis supported DX10, for god's sake. DX9 is outmoded.

    2. There are shitty textures everywhere.

    3. The graphics options are incredibly limited, and changing them does not produce an appreciable difference in visual quality. Seriously, the lowest preset looks about as good as the highest. Turning the setting up only adds some shader anti-aliasing and pours on some ridiculous motion blur effects.

    4. The blur effects are terrible. There is so much motion blur; it quickly becomes nauseating. The blur on the light sources is very unrealistic. It's almost as if Crytek knew that the graphics were going to be bad, so they attempted to hide the unsavory reality using excessive blur.

    5. The faces look really bad.

    6. The lighting is fake and overdone, just like the motion blur.

    7. Environments are no longer destructible.

    8. Object deformation effects are quite lackluster.

    In my opinion, the game does not look good at all; it is a big step backward from the original Crysis, and it is absolutely smashed by true modern contenders like Just Cause 2, Metro 2033, and Battlefield Bad Company 2.

    Does anyone care to reopen the Crysis 2 vs Metro 2033 argument now?
     
  7. red1776 Omnipotent Ruler of the Universe Posts: 5,796   +24

    I know what you mean, why bother? its not even close. The textures in Crysis 2 are absolutely horrendous. Not to mention, this isn't even a corridor shooter...its a phone booth shooter.
  8. dedparrot Newcomer, in training Posts: 67

    This will be of use to any users who want a really simple way to change individual graphics settings in Crysis 2.

    It may also be of help to the Techspot author who may want to tweak settings in his game.

    As of yet, the application is in version 1.7, and Crysis 2 players are super appreciative of it.

    http://forums.steampowered.com/forums/showthread.php?t=1807934
  9. Zecias TechSpot Booster Posts: 202

    it really depends. if you are used to playing all games at around 30 fps and crysis doesn't ever go beyond 30-40 then you probably won't notice any lag(playing games at a constant fps makes them seem less laggy).
    if you usually play games around 60 fps and can't tell the difference between 30-60 fps, then i don't believe you ._.
    i'm used to playing at 60 fps, so when i play at 30 fps, it is very noticeable, especially when the fps jumps between 30-60.
    is it so hard to believe that people can see lag after 30 fps? after all, the human eye can see well past 30 fps.
    is it so hard to believe that people notice the difference between 30 and 60 fps?
  10. Arris TechSpot Evangelist Posts: 4,308   +17

    Steve, I dunno what I was thinking. 4GHz - 4.6GHz makes little if no difference to Crysis2, and even the AvP benchmark with crossfire enabled benefits very little from the extra 600MHz. Think I must have been confused between the CPU overclocking and when I was overclocking the GPUs slightly. Probably just overcome with the general performance increase in just about everything else when I moved to this new system.

    I get the flickering with 2 cards enabled. Must be something in the code that dislikes even numbers of cards :D
  11. Steve TechSpot Staff Posts: 875   +67

    If your standards are low that does not make the opinion of others wrong. And in my opinion 30fps feels very laggy.

    I am aware of Wasdie’s program and have used it, trouble is most of the tweakable settings break the game. Enabling MSAA for example destroys the image quality, makes the game look very washed out.

    Hi Arris, yeah I have done huge amounts of CPU scaling testing in my time and you will find that at 4GHz there is literally no game that can max out a Core i7 processor.

    Why are you spamming this thread with driver links? Please don’t…

    You do realize that the focus of this article is GPU and CPU performance, not to make wild assumptions?
  12. I'm not sure if I trust these benchmarks.

    I have an XFX 6870(OC to 950/1250) and I usually hover around 60 fps but the lowest I've seen it drop to is 38 fps. This is on Extreme settings with V sync on at 1920x1080.

    Not sure if its my computers specs that are making a difference?
    965BE @ 4ghz
    8gb Corsair XMS3 1333mhz
    Gigabyte 890GPA-UD3H

    I don't believe I'm imagining things either.
  13. Technochicken TechSpot Paladin Posts: 900

    Nice review. Judging by the given CPU results, and the framerates I got in the multiplayer demo, it looks like my system was more CPU than GPU bottlenecked. And I see TechSpot has a new bench system!
  14. glad and somewhat shocked to see how hyper threading helps core i3 compared to a higher clocked e8500.so core i3 budget users are not to be dissappointed
  15. Arris TechSpot Evangelist Posts: 4,308   +17

    The benchmarks are of 1920x1200, you are running 1920x1080, so a few hundred thousand pixels less processing per frame. Are you using the same Catalyst version used for the testing? It's possible you have a better performing video driver.

    There are many things that can influence your own results. You even mention that you have your card overclocked. Unless you are going to test at default clock and memory at same resolution with the same drivers it's pretty pointless to state your impression of what your cards performance is like and say you don't trust the benchmarks listed here. I too thought my card on my system was performing better than the benchmarks but really it wasn't. Also remember the frames per second listed are also an average.One drop to low fps can really drop your average.
  16. Julio Franco TechSpot Editor Posts: 6,055   +121

  17. I can't believe you even let them off the hook for this consoled up garbage.

    I'm not reading this rag of a magazine anymore.
  18. I think you need to try the RadeoPro "fix" for crossfire.

    I'm running 2x6970 at 1920x1200 with much more than Extreme settings (See Nebby's mods on the Cry2 thread at B3D) and I'm pegged at 60 fps with vsync on (60 Hz screen and proccy is 2600k at 4.5GHz). These numbers are much higher than reported here.

    To do this I am using RadeonPro to force the Bioshock Crossfire profile and D3D Triple-buffering.
    My guess is that this modification would up the 6990 results.
  19. Per Hansson TS Server Guru Posts: 1,796   +66

    Oh I hear ya.
    I noticed the poor textures first thing I did when playing the leaked version at a friends house for roughly 10 minutes.
    Then I chalked it up to being a leaked beta, and that the full version would use high resolution textures.
    But nooo, it's very easy to see for yourself, in the first mission you go up to a wooden door, the first one is open a little so you might not notice it, but then you enter an area and kill 3 guys, after that is the same door again, but now locked.
    So you go up to it and press "use" (since it's the logical thing to do, and yea, my screenshot below really tells it all.)
    This is at 1600x1200 "Very High" quality
    In the Dosbox Window I am playing Duke Nukem 3D, which looks better to you?

    http://www.techspot.com/gallery/fragfest26-gaming-in-general/p4238-crysis-2-vs-duke-nukem-3d.html
  20. I do not understand how you people got theese results nor how you came to the conclusion that this DX9 bull is as demmanding than the original Crysis which you did say?

    Which in FACT on my desk tested with 4 video differnet video cards it is not as demmanding as the original and that was the whole point this tim around !!!!

    The original was meant to be a bear that few then current cards could even play, while this round Is basically a console port they are going to tweak along the way, and this was done so on purpose to SELL, SELL, SELL, to the masses and SELL it has on XBOX 360 and PS3.

    From my own testing and many sites on the web such as Toms Hardware, crysis 2 is in no way as demmanding as the original. My HD 5770 struggled with the original across the board, yet it plays Crysis 2 quite nicely. Also the results you got with AMD 6xxxx cards puzzles me and defies what I have on my desk.

    My P2 1055t overclocked mildly to 3.73 with an HD 6950 and it laughs at this game easily getting 60 FPS on Ultra Super Dooper high or whatever they call it today!