• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Does SLI scale well with crysis yet ???

Seems a tetchy little thread this one! :)

The only benchies I can still find when I had my 8800GT setup are the following;

1440 x 900 - Windows XP - Very High Settings (DX9 Hack)

One 8800GT

Min = 9

Max = 29

Avg = 24


Two 8800GT SLI

Min = 7

Max = 53

Avg = 35


Might give you some idea of what to expect, but it definately works.

That was with an E2180 CPU at 3Ghz on a Gigabyte N650 SLI Motherboard.

HTH :)
 
Quads do not help with Crysis, all that happens is the 2 threads it uses get split across the four cores.
A while ago i swapped out my e6700 with a q6600 running the same multi and FSB expecting to see gains and there weren't any.

However if someone posts some proof that quads are of benefit in Crysis i will go to the trouble of re examining this.

To answer Jiggers question.

Settings for drivers are defaults i don't mess with the performance settings.
Forceware 169.21
FPS figures for Crysis were obtained with settings described in the Crysis benchmark thread.
Crysis ini files are untouched.
 
Last edited:
PCZ said:
Quads do not help with Crysis, all that happens is the 2 threads it uses get split across the four cores.
Click to expand...
No it uses all four cores simultaneously as well, but as the game is so GPU limited you have to get it in the right situation to make it properly utilise all four at once otherwise the CPU is motly left alone.

PCZ said:
However if someone posts some proof that quads are of benefit in Crysis i will go to the trouble of re examining this.
Click to expand...
I've done this a number of times and you can do it yourself too. Just run the CPU benchmark in the following settings: 800x600, all settings low, physics high or very high.

This forces the benchmark into an entirely CPU-bound state.

Edit: To save you the leg-work, I've dug up the old screenshot I took: -

82wr1c7.png


As you can see parts of this image clearly show all four cores being used at once, mostly three though but since it's a physics benchmark and doesn't have to deal with any AI then it's actually stressed less than in a real-world gaming situation. The only difference is that the resolution drop has forced the effects on the CPU to be more prominent.
 
Last edited:
Sorry dude, that picture doesn't prove anything. If I run World of warcraft or any single core game for that matter you see the load shared between 2-3 core not just one. Its the operating system that schedules the processed to the individual cores, not just running them on one.
 
You need to show a total of a least 2+ cores being maxed. I suggest for testing down clocking your four cores as much as possible and doing the physics test at a really low res but I think 2 cores only is correct
 
PCZ said:
Please people stop spewing FUD
Click to expand...
PCZ, if your findings are different from other forums members then you should find a better approach than to call them all ******** :D

Seriously, tone the attitude down abit and try and work out how other people using SLI set-ups are not getting as good a result in the CRYSIS bench as you!
 
My TriSLI rig runs much better than 1 card, 2 is an improvement and 3 takes it further still, completely playable on max settings @ 1680-1050 dx10

I dont know why people slate SLI so much from my experience it works well. I sometimes wonder if most of the people slating it have even tried it themselfs.
 
Its not proof of the game using 4 cores just that windows will display all 4 cores being in use.
Try comparing it to rmclock or some other core usage graph util
 
You must log in or register to reply here.
Back
Top Bottom