Crysis turns up nose at quad

[D!Z]

Originally Posted by ynnek

Who is this guy?

Okay, I can perhaps imagine, someone from a relatively big publication getting a prerelease gold copy of the game from Crysis. But if I were EA, I'd put him under a NDA to not say anything until after the game is released...... Anyone else think this is odd? Or this guy Peckham and PCWorld thought it was worth breaking the NDA for...... Or the EA didn't make him go under NDA, or the NDA expired already...

I think hes a bit full of it. A x6800 for 400.00? maybe ebay, on newegg they are almost $1,000.

[[EOCF] Tim]

Originally Posted by D!Z

I think hes a bit full of it. A x6800 for 400.00? maybe ebay, on newegg they are almost $1,000.

£400 perhaps? Anyway, Crysis did seem to load all 4 cores, not much, but it's definitely doing something on 3 and 4.

[miahallen]

Originally Posted by StellaArtois

Good testing. Looks pretty limited at 2 cores and above. Could you please try a run with all crysis settings to low, except physics at very high?

CPU @ 333x9=3000MHz
PCIe @ 100MHz
RAM @ 800MHz 4-4-4-12 T2

1 Core
0.09/17.46/45.47

2 Cores native
20.18/36.12/47.47

2 Cores non-native
22.11/36.98/47.24

3 Cores
19.87/36.38/48.76

4 Cores
20.36/36.28/46.31

CPU @ 333x9=3000MHz
PCIe @ 100MHz
RAM @ 800MHz 4-4-4-12 T2
Only Physics @ VH

1 Core
0.50/13.98/77.53

2 Cores native
46.11/69.23/86.12

2 Cores non-native
24.55/67.64/89.19

3 Cores
47.47/69.65/88.47

4 Cores
46.66/70.32/88.50

In both cases the difference was about 1% in average framesrate between dual core (native) and quad core. I think the quad core myth is officially busted - for the demo.

[miahallen]

Just for the record, I think all of my single core results should not be compared directly, I think they are bugged. I would get massive pauses (2-3 seconds where the game screen would go black, and the task manager would tell me crysis.exe was not responding. This would occur once every 10-15 seconds during each benchmark run. I think the game was seeing multiple cores and trying to use them, but becasue I had manually changed the affinity, it was having issues. This only occured during the single core tests however.

[Mr_LoL]

Yay Dual Core ftw!!!

[StellaArtois]

Originally Posted by miahallen

CPU @ 333x9=3000MHz
PCIe @ 100MHz
RAM @ 800MHz 4-4-4-12 T2
Only Physics @ VH

1 Core
0.50/13.98/77.53

2 Cores native
46.11/69.23/86.12

2 Cores non-native
24.55/67.64/89.19

3 Cores
47.47/69.65/88.47

4 Cores
46.66/70.32/88.50

In both cases the difference was about 1% in average framesrate between dual core (native) and quad core. I think the quad core myth is officially busted - for the demo.

Thanks! Looks conclusive...

[[EOCF] Tim]

Definitely using 4 cores. Q6600 @ 3.6Ghz.

All at high, except water, sound, and textures at very high.