Far Cry benches: q9550 vs. i7-3820

[magellan]

My old q9550 system died but not before I managed to squeak out a Far Cry Matto3 benchmark w/my 6970 using the special 64-bit version of the game. This was with my q9550 @ 503 * 8.5 = 4275Mhz and the same Gskill memory I'm using in my i7-3820:

6950 flashed as 6970 @ 1001/1481, MSAA, 16xAF high quality, vsync, 3x temporal AA enabled at 45 FPS, texture LOD -10, flip queue size 5 EQAA enabled, morphological enabled, 4xAA edge detect,

=====================================
TimeDemo Play Started , (Total Frames: 13299, Recorded Time: 366.38s)
!TimeDemo Run 0 Finished.
Play Time: 259.18s, Average FPS: 51.31
Min FPS: 27.59 at frame 1756, Max FPS: 72.36 at frame
11351
Average Tri/Sec: 8197026, Tri/Frame: 159752
Recorded/Played Tris ratio: 0.71

Just because I'm doing some benchmarks before installing a new vid-card I performed the same benchmark on my
i7-3820:

Far Cry Windoze 7 64-bit
i7-3820 @ 128.5 * 37 = 4754.5 Mhz, DDR3-2056
6970 @ 1001/1481, 4xAA MSAA, edge detect, EQAA enabled , morphological enabled, 16xAF high quality, vsync, 3x temporal AA enabled @ 45 FPS, texture LOD -10, flip queue size 5

VRAM usage:
dedicated 599 MiB
dynamic 79 MiB
total 678 MiB

=====================================
TimeDemo Play Started , (Total Frames: 13299, Recorded Time: 366.38s)
!TimeDemo Run 0 Finished.
Play Time: 158.91s, Average FPS: 83.69
Min FPS: 37.06 at frame 10562, Max FPS: 116.23 at
frame 6125
Average Tri/Sec: 13168221, Tri/Frame: 157347
Recorded/Played Tris ratio: 0.72

I didn't really expect to see any difference at all.

 

[doz]

People fail to realize how old the C2Duo/Quad architecture really is. For some, its fine, but for gaming its VERY far behind. Some people still claim their C2Quad @ 3.8/4.0ghz will keep up with a 2500k/3750k/4570k in GAMING (not benchmarks) at the same clocks.

Glad to see your results. Its what Ive been saying all along. I noticed a good jump too (about 15-40% depending upon the game) from my i7-920 @ 4.5ghz to a 2600k @ 4.9ghz. Granted a little faster clockspeed, but not enough to warrant the drastic jump as some people claim there isnt much difference for gaming between a 920 and 2600k.

 

[magellan]

For some reason, I can't run the Far Cry 64-bit, Matto3 benchmark at anything higher than 4xSSAA. If I run it at 8xSSAA it always crashes.

Of course Far Cry is DX9 and old tech, so it's not going to be GPU-bound unless you crank up the AA and use super-sampling. At 8xSSAA my 6970 has noticeable stutters but still uses negligible amounts of VRAM (not even 1 GiB). I might try turning off compressed textures to see what happens.

 

[EarthDog]

Thanks for posting...however a better, more accurate and conclusive test would be to run the CPUs at the same clockspeed. As it stands, the newer chip has a 500mhz lead. Also the settings are different between the two runs? One cant compare apple's to apple's with different settings and speeds. Why is vsync enabled? In benchmarking vsync should never be used as it limits fps (counterintuitive when comparing fps ). What resolution was this tested at?

I like the thought, but the implementation leaves a lot to be desired to get meaningful results between the two CPUs. The story will remain similar, but to compare the two cpus, settings need to be the EXACT same for a valid comparison, and as I read your post, I think there are some differences and concerns, lol!

 

[magellan]

Thanks for posting...however a better, more accurate and conclusive test would be to run the CPUs at the same clockspeed. As it stands, the newer chip has a 500mhz lead. Also the settings are different between the two runs? One cant compare apple's to apple's with different settings and speeds. Why is vsync enabled? In benchmarking vsync should never be used as it limits fps (counterintuitive when comparing fps ). What resolution was this tested at?

I like the thought, but the implementation leaves a lot to be desired to get meaningful results between the two CPUs. The story will remain similar, but to compare the two cpus, settings need to be the EXACT same for a valid comparison, and as I read your post, I think there are some differences and concerns, lol!

All the benchmarks were run at 1600x1200. I used vsync because it's required for the use of temporal AA.

I was really running these benchmarks to see what kind of difference my incoming GTX 780 will have over my overclocked 6970.