PCI Express 2.0 vs 3.0 GPU Gaming Performance Review

Wondering about upgrading to the new Ivy Bridge CPU and PCI Express 3.0 platform? Curious to know if you'll be gaining or losing performance? We compare single, dual-GPU, triple-GPU, single, and multiple display configurations on Ivy Bridge PCIe 3.0 and Sandy Bridge on PCIe 2.0 platforms.

continued...

Highest Playable Settings - GeForce GTX 680 2-Way SLI

All of the settings below are what we found as the highest playable and best overall settings for GeForce GTX 680 2-Way SLI on Sandy Bridge PCIe 2.0 and Ivy Bridge PCIe 3.0. We ran GTX 680 SLI on three Dell 2408WFP displays to find highest playable setting on triple-display at 5760x1200.


Max Payne 3

Article Image

Article Image

(Click Graph for Larger Image)

In Max Payne 3 we were able to play this game at the highest possible in-game settings at 5760x1200 on both systems. We found that 5760x1200 worked well with all the in-game settings at the highest values, including tessellation and ambient occlusion. We did have to use FXAA, as 2X MSAA was just a bit too much of a drag on performance to be playable.

Performance-wise we experienced exactly what we did with a single GTX 680, that is that PCIe 3.0 system showed technically better framerates. With a 72.0 FPS average the Ivy Bridge PCIe 3.0 system was 5.572% faster than the Sandy Bridge PCIe 2.0 system. We are looking at framerates that are averaging in the upper 60 FPS and lower 70 FPS range with 2-Way SLI at this resolution. Again, a 5.572% performance difference doesn't equate to a better gameplay experience. Still it is interesting that it is consistent with the single card results.


Battlefield 3

Article Image

Article Image

(Click Graph for Larger Image)

With Battlefield 3 we were also able to achieve the same gameplay settings on both systems. We found that 5760x1200 with 2X MSAA plus FXAA was playable with all Ultra in-game settings.

On the performance front we have a bit of a reversal. Here we are seeing the Ivy Bridge PCIe 3.0 system average 53.8 FPS, and the Sandy Bridge PCIe 2.0 system average 54.1 FPS. This is a mere 0.3 FPS difference or 0.555% decrease in framerate. This fraction of a percent is well within the margin of error for testing, so in effect both systems performed "identically" as we saw previously in single card GTX 680 testing with this particular title.


Batman Arkham City

Article Image

Article Image

(Click Graph for Larger Image)

In Batman: Arkham City we once again found that we could play at the same in-game settings. We found that 5760x1200 was playable and that we could even enable 4X MSAA with 2-Way GTX 680 SLI. This is with all the in-game settings at the highest values including tessellation and shadowing. We were also able to have PhysX turned on the Normal level with 4X MSAA. If you opted to instead use FXAA you could then enable High PhysX.

In terms of performance we found once again identical gameplay performance between Sandy Bridge PCIe 2.0 and Ivy Bridge PCIe 3.0. The Ivy Bridge system showed us a 1.333% framerate increase.


Skyrim

Article Image

Article Image

(Click Graph for Larger Image)

With Skyrim we once again found that this game was playable at the same exact settings on both systems. We found that 5760x1200 was playable with 8X MSAA plus FXAA plus 4X TR SSAA enabled at this high resolution. The gameplay experience was the same.

In terms of performance, like BF3 it looks like PCIe 2.0 system is faster, but we are talking a decrease of 2.059% in framerate. We can easily again consider this small difference within the margin of error. Skyrim is a very dynamic game, and certain things like NPC pathing change between run-throughs, so this is normal. Interestingly though, the minimum FPS is higher on PCIe 3.0, that 0FPS on PCIe 2.0 indicates a load screen, and it didn't drop to 0 on PCIe 3.0 like we are used to seeing in our past Sandy Bridge system testing.


The Witcher 2

Article Image

Article Image

(Click Graph for Larger Image)

In The Witcher 2 we found that we could play at the same settings on both systems. At 5760x1200 this turned out to be all Ultra settings except Ubersampling turned off.

In terms of performance we did get faster framerate performance with the Ivy Bridge PCIe 3.0 system. The exact performance increase is 8.421% over our PCIe 2.0 system. However even with 8.421% better framerate performance this doesn't improve the gameplay experience or let us increase more visual fidelity settings.


Summary

In summary, there were two games once again that intrigued us as far as performance advantage while the others showed virtually no differences in even framerate, much less gameplay experience. In Max Payne 3 we received a 5.572% increase to framerate performance with the Ivy Bridge PCIe 3.0 system. In The Witcher 2 we saw an 8.421% increase to framerate performance with the PCIe 3.0 Ivy Bridge system. Again however we gained no benefits to the actual real-world gameplay experience in any way at all.