Oooh god I can't wait for my 9262 to ship.. I'm having to borrow people's computers because I sold mine like an idiot before I got this one in my hands.. and all you people are running Crysis... agghh!!
-
-
what resolution and what is your average fps in benchmark?
-
Magnus is playing at 1920x1200.
-
My average with that config and modified Cvars is 31 fps running it in DX9. I think it might even be faster in XP actually. I only have Vista, but with Vista I have 31 fps. This with GPU´s overclocked to 625/1500/950, eleron I change my clocks all the time on the GPU´s
Those with extreme CPU´s or even Quads should probably see even more boost in FPS. But to me this config and Cvars looks the best and plays much better and looks like DX10 Very High. In fact I changed all Cvars to include all Very High. -
magnus, can you email me your modified cvars zipped up?
-
Yes sure I will do that when I get home from work, PM me your mail address, same with you eleron easier that way.
-
ok got your modified teaks and cvar group. there is a definate improvement in fps. i have not fully examined what this mod changed but it appears that it added some dx10 effects like god rays and changed the lighting around at the expense of reduction in visuals for very far objects. it may be adouble edge sword but it does look very nice, indeed. i would say that close quarters combat this mod makes things look nicer but viewing a wide open field from afar has reduced quality.
anyway, i ran the test again with this mod at wuxga (1920x1200) and got 27.42 average fps compared to 21.44 with non-modified tweaks and only the dx10 very high hack.
for some reason this mod does keep in memory when i change the resolution so when i exit the game it defaults back to wuxga. i cannot run tests at lower res because of this. i will look into it when i have time.
i think it is a very nice mod overall and allows very high settings at higher frame rates, in general. -
So ARGH, just to get this straight, do you think the system in my signature can handle Crysis @ very high at 1400x900? What I mean is, will there be times when the framerates become unplayable? (many characters on-screen, huge explosions, etc). I saw the minimum framerates yes, but I'm just curious, was it jerky? thanks again btw, you've given the best data on this I've seen so far!
-
1440x900 all very high dx10 = 28.87 fps avg. sli card
that is with my 2.4ghz cpu. if you get a faster cpu, like a 2.8 or 3.0 or better you will be in the low 30's average fps and that, to me, will be acceptable where it will be smooth enough to play the game overall. having said that, yes, there will be times where you will reduce the very high setting to high in that resolution after you soaked in the eye-candy and just want to play the particular episode smoother.
1440x900 all high dx9 = 38.56 fps avg. sli card
that is with high settings. i found myself playing in this resolution quite often since i craved better framerate when the rough spots kicked in. the 10 fps boost really does make a difference.
this will all depend on just how picky you are about frame rates. imo, 30 fps average within the game's built in benchmark tool is the absolute minimum. when the game starts out it runs very smooth but after that it tends to throw some high intensity scenarios at you which will have you thinking about lowering settings and or resolution to get through it. some later levels play just as smooth as the first level. it's those tight frantic situations that will give you the fps hurt.
also, and again depending on how picky you are, you can use the modified tweak that magnus uses and it will give you the very high setting and more with about as good framerates as high setting.
the bottom line is that yes, the system you are looking at will be able to handle very high settings but it will take some adjustment on your part to figure out where you want to be....althought it appears i have done all the work for you lol.. -
ARGH I have set 1920x1200 as default res in the autoexec.cfg. You can remove those lines so it doesn´t always revert back to 1920x1200.
-
well, I do have the 9450, and I have almost 300 mhz over yours, plus a bit more in the cache. My 2 9800GTs are unfortunately basically yours. How much RAM you got? Do you have the 667 or the 800?
-
ok i have completed the testing with the modified version! i will re-post all the benchmarks again to make things easier.
crysis internal benchmark utility full screen mode Benchmark_GPU batch file;
800x600 all high dx9 = 49.16 fps avg. single gpu
1280x960 all high dx9 = 31.91 fps avg. single gpu
1440x900 all high dx9 = 29.62 fps avg. single gpu
1680x1050 all high dx9 = 23.37 fps avg. single gpu
1920x1200 all high dx9 = 18.86 fps avg. single gpu
800x600 all high dx9 = 43.50 fps avg. sli enabled
1280x960 all high dx9 = 41.67 fps avg. sli enabled
1440x900 all high dx9 = 38.56 fps avg. sli enabled
1680x1050 all high dx9 = 35.19 fps avg. sli enabled
1920x1200 all high dx9 = 30.28 fps avg. sli enabled
--1280x960 all very high dx10 xp hack = 30.75 fps avg. sli enabled
--1440x900 all very high dx10 xp hack = 28.87 fps avg. sli enabled
--1680x1050 all very high dx10 xp hack = 25.24 fps avg. sli enabled
--1920x1200 all very high dx10 xp hack = 21.44 fps avg. sli enabled
**1280x960 all very high dx10 xp hack with modified tweaks = 37.42 fps avg. sli enabled
**1440x900 all very high dx10 xp hack with modified tweaks = 34.37 fps avg. sli enabled
**1680x1050 all very high dx10 xp hack with modified tweaks = 32.18 fps avg. sli enabled
**1920x1200 all very high dx10 xp hack with modified tweaks = 27.42 fps avg. sli enabled -
-
Glad you like it ARGH and a nice FPS boost too for you.
-
magnus, i am curious to know what your scores would be with your m1730. it would be a nice comparison.
-
how much difference is there playing at 1680x1050 vs 1440x900 or 1280x960 in terms of how good the game looks? Usually the first thing i do when i adjust settings to get better performance is to turn down the resolution, seems like its a huge deal to you guys.
-
it impacts the visual quality quite a bit because of the native resolution of the lcd panel. the lower the resolution from it's native the more blurry it looks. you can avoid this issue if you play the game in windowed mode, but SLI does not work like that. the game must be in full screen mode for SLI to kick in, unfortunately.
one way around this is to hook up the 9262 to an hdtv, like the 720p 55" plasma that i have via dvi to hdmi connector. then you can play the game in 1280x720 and have the best fps. -
Or maybe I could figure out how to hook it up to this: http://www.myvu.com/Crystal.html -
ARGH this is my best average so far running in DX9 with this config at 1920x1200.
!TimeDemo Run 2 Finished.
Play Time: 61.48s, Average FPS: 32.53
Min FPS: 18.50 at frame 1952, Max FPS: 47.36 at frame 994
Average Tri/Sec: -13958181, Tri/Frame: -429085
Recorded/Played Tris ratio: -2.14 -
Petrov. -
-
I tried it on mine, dx10 though, started stuttering 5-7 fps! -
-
-
Yes with my GPU overclock, currently 625/1500/950
NP5796 Crysis Performance
Discussion in 'Sager and Clevo' started by h575309, Jul 31, 2008.