Crysis 3 performance

[Name]

on my GTX 670 i received 50-60 fps on high settings medium AA and vsync on the first level, the rain tanked my fps right down to 30-35 though.

Moved it down to Medium setting with AA off, still looks very good, received anywhere from 90-140 fps in game on the same level, much better performance

 

['Wes]

I have a good 3930 OC to 4.8 and a OC 7950. So meaty 6 core processor with 32GB RAM

All of my settings are on Very high with MSAA set up. Not sure on frame rates, but certainly playable on solo campaign. Playing on 1920x1080 on my Asus IPS monitor.

I think the game benefits from more cores, my 3570k was at 100% usage when i checked the other night, just like Crysis 2 liked 8GB RAM.

 

[JoeF1]

I think the game benefits from more cores, my 3570k was at 100% usage when i checked the other night.

Yeah i read somewere that more cores the better as it nails the CPU. Dont quote me on that though.

Runs so well on my standard rig (darnt over clock yet, too new). Have the GFX on very high apart from the water, shaders and shadows on high. I have the TXAA on 1 aswell and vsync off - followed the tips you guys have left. Can i ask what the difference is with MSAA on/off?

Dont know the FPS because i dont know how to check, is their just a console command like BF3?

Thanks.

 

['Wes]

Just performance, i didn't notice a difference between MSAA, SMAA, FSAA and turning AA off altogether visually, the only difference was TXAA visually, as it adds a blur(ish) effect.

MSAA knocked the FPS off by about 15.
TXAA knocked it off around 10-12.
SMAA knocked it off around 5-10.
FXAA none at all.

Here's two screenies of me running with everything maxed with TXAA High but Shadows, Objects and Shading on high.

FPS was 35+, but dipped to 30 on occasion - playable frame-rates for me.

 

[Name]

What card are you using Wes?

I received similar fps on my card on high settings, medium AA (GTX670)

 

['Wes]

What card are you using Wes?

I received similar fps on my card on high settings, medium AA (GTX670)

KFA2 660 OC, i game on a 20" Monitor with 1600x900 resolution.

 

[Name]

Ah i was about to question the similar performance then against my 670 then.

Realised you're playing 1600 X 900, i might try that and see if it improves performance, only with certain games though.

 

[Guest2]

It definately will. Like going from 1440p to 1080p (very high) goes from 29fps to 47fps

It looks horrible in 1080p on a 1440p native screen though. I play now in 1440p, all very high but shadows, shaders and something else in high

 

[bakes0310]

So without reading through the whole thread whats the reason with people with high spec systems not coping to play this game at max settings?
Is it the game or even high end rigs not ready for the game yet?

 

['Wes]

So without reading through the whole thread whats the reason with people with high spec systems not coping to play this game at max settings?
Is it the game or even high end rigs not ready for the game yet?

Not sure, the only thing i've noticed causing my system problems was my 3570K at 4.4GHz basically hitting 90-100% usage, which i found odd, and my GPU getting nowhere near those numbers.

 

[Rusty0611]

Not sure, the only thing i've noticed causing my system problems was my 3570K at 4.4GHz basically hitting 90-100% usage, which i found odd, and my GPU getting nowhere near those numbers.

Seems like there's a CPU bottleneck for you then. Never seen one that obvious on an overclocked i5 2500k / 3570k though. Normally it's more subtle like on BF3 64MP.

I've got an i7 3770k at 4.4 GHz so I will get a copy soon and see if the extra 4 threads help.

 

[wolvers]

3770K is only slightly faster than the 3570k.

Looks to me like it doesn't support HT and the slightly higher clock speed has put it ahead.

 

[Rusty0611]

3770K is only slightly faster than the 3570k.

http://i1056.photobucket.com/albums/t364/Hunbug76/Crysis3/Crysis-3-Test-CPUs-VH-720p_zps6d1a466c.png[IMG]

Looks to me like it doesn't support HT and the slightly higher clock speed has put it ahead.[/QUOTE]

What, 100 MHz has boosted average FPS by 12.3% and minimum FPS by 6.8%?

No way is that just down to 100 MHz lol.

 

[MOOGLEYS]

I see no difference at 1440p with a 2500k @ 4.6ghz and a 7970 @ 1200/1600 or a 3770K @ 4.6ghz and a 690 GTX at stock.

I think a lot of the performance issues are more to do with the poorly optimised engine rather than Pc grunt.

 

[wolvers]

Yeah, you're right I didn't really looks at the numbers.

Still, I don't think HT is supported. In this review the CPU loads of the Intels show this;

http://gamegpu.ru/action-/-fps-/-tps/crysis-3-test-gpu.html

Edit; the graphs are easier to see together here;

http://www.overclock.net/t/1362591/gamegpu-crysis-3-final-gpu-cpu-scaling

Maybe AMD have gimped HT, seen as this is a GE title. Did HT work in Crysis2?

 

[Rusty0611]

Yeah, you're right I didn't really looks at the numbers.

Still, I don't think HT is supported. In this review the CPU loads of the Intels show this;

http://gamegpu.ru/action-/-fps-/-tps/crysis-3-test-gpu.html

Edit; the graphs are easier to see together here;

http://www.overclock.net/t/1362591/gamegpu-crysis-3-final-gpu-cpu-scaling

Maybe AMD have gimped HT, seen as this is a GE title. Did HT work in Crysis2?

Can't remember mate.

Yeah interesting results. Definitely poses more questions than answers.

I see no difference at 1440p with a 2500k @ 4.6ghz and a 7970 @ 1200/1600 or a 3770K @ 4.6ghz and a 690 GTX at stock.

I think a lot of the performance issues are more to do with the poorly optimised engine rather than Pc grunt.

Seeing a difference is different to whether there is a difference or not although I agree that if you're buying something then it becomes a consideration. You're right, it could be engine problems but I'll have a look at my task manager when running it to see what happens.

 

[wolvers]

Are there any other cases of HT not being utilised, but more cores are? I always assumed that they were used as extra cores, but that they didn't scale to the same performance as physical cores. I'm sure there's someone here that can explain it better.

 

[fantasticCG]

Just ran a couple of benches using fraps for a minute from when the doors open to the jungle to the cut scene with the view and turret towers.
On the dell 27" 1440p using 2x680s latest drivers and using EVGA PX
PT132%
GPU +175mhz
MEM +250mhz
Resulting in a 1307mhz on card 1 and 1127mhz on card 2

Vsync on
MSAAx4 Min 36 Max 61 Avg 50.783
MSAAx2 Min 56 Max 56 Avg 56
FXAA Min 52 Max 62 Avg 59.533
1xSMAA Low Min 53 Max 62 Avg 59.833
Didnt bother using the others the TXAA drop was around 37fps and lower sometimes.
The in game settings where from the geforce experience program.

AF 16x
AA= different ones used. it suggested 1x SMAA Low
FS= Yes
Game effects= Very high
Objects= Medium
Particles= very high
Postprocessing= high
shading= high
Shadows= medium
Texture res= very high
water= very high

 

[porkmince]

optimised for 1440p and GTX 680 SLI

with my rig : having been playing around for some time and reading the threads here

@ 1440p I have settled on running everything very high - no motion blur and no v-sync ( using NVidia inspector to limit to 60 FPS) and adaptive v-sync in nvidia panel

using no AA - at 1440p I really cant see any jaggies - switching FXAA takes a slight hit and I prefer a smoother FPS

textures very high and its running very smooth ( buttery) and really enjoying the game .

 

[Ste]

Playing on a GTX670 and i7 920 @ 4.2ghz (HT on) by CPU usage seems to be around 25-40%. Doesn't sound right compared to people on here talking about 100% all the time. Any ideas? I'm playing maxed out at 1080p, 1440p doesn't make any difference, not that it would since you're just emphasizing the GPU weak link...