Benchmarks & Conclusion
Benchmarks & Performance Analysis
GeForce 368.39 vs. the GTX 1080/1070’s launch drivers and vs. GeForce 368.22
Here are our results of 26 games and 1 synthetic test. The drivers are noted on the chart. “X” means the benchmark was not run and “NA” means the benchmark could not be run or completed. And an Asterisk (*) means that there are some real issues with the visuals.
All results except for Firestrike are expressed in average fps (in Bold) and also generally in minimum fps (where they make sense, in smaller italics next to the averages). Each set of drivers are compared against the other in the adjoining results column in white (newest driver) versus black (older drivers).
We can see many performance gains for our 3 test cards with the newest drivers in many games. It appears that Nvidia’s driver team is hard at work tuning the performance of the GTX 1080 and the GTX 1070 as befits brand new architecture, with generally positive performance increases. The negatives are usually quite small and do not practically affect playability. The only real outlier we noted is with Shadows of Mordor which shows a significant decrease in performance for the GTX 1080, and we suspect that default Ultra was originally used instead of our current maximum settings.
The GTX 1070 was released later than the GTX 1080 and shows less change although it still gains decent performance with the latest games. Its only outlier is Fallout 4 which showed a small performance decrease although it should not affect playability. And the GTX 980 Ti has not been neglected as the new driver generally shows incremental performance increases as befits older architecture.
This GeForce 368.39 driver is very stable, and in our opinion, a very worthwhile upgrade. We saw only a few cases with slight decreases that did not affect these few games’ playability. We are quite impressed with GeForce 368.39 just as we were with with GeForce 368.22 last month.
Finally, let’s check out our bigger picture next.
The Big Picture
The Big Picture, once reserved solely for our video card reviews, has been recently expanded to feature all of our competing top video cards on recent drivers. Driver revisions for each card are noted on the chart. An “X” means the benchmark was not run, while “NA” means the benchmark could not be run or completed. And an Asterisk (*) means that there are some real issues with the visuals. All results except for Firestrike are expressed in average fps (in Bold) and also generally in minimum fps (where they make sense, in smaller italics next to the average results).
Let’s head to our conclusion.
We would recommend upgrading to the latest WHQL GeForce 368.39 drivers as they have brought some performance improvements to many of our tested games, especially to the latest ones. Overall, there are generally incremental advantages over older drivers, but with also some rather large performance increases, particularly for the GTX 1080 since its launch drivers. We always recommend using Nvidia’s latest WHQL driver for Windows 10.
GeForce WHQL 368.39 is generally the best driver to use for the latest really fun games that we are playing, including Rise of the Tomb Raider, Far Cry Primal, Hitman, Tom’s Clancy’s The Division, DOOM, and Mirror’s Edge Catalyst.
It appears that both Nvidia and AMD intend to release new drivers with each new major PC game release, and we are looking forward to our next evaluation. AMD has released their own June Crimson Software driver, but it is another Beta Hotfix non-WHQL driver which gives real issues for users with Blu-Ray and other DRM programs. It has been over 2 months since AMD has released a WHQL driver, but we will also chart AMD’s latest driver performance next week.
In the meantime, join BTR’s Community for great tech discussions and also feel free to comment on this Performance Analysis in the comments section below.