Crysis turns up nose at quad

[miahallen]

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]

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.