Should I update Nvidia GPU drivers? Benchmarks by version
This article lists benchmarks with fps and frame times on all the Nvidia drivers of the past six months. Find out if it is worth updating your driver version.
When Nvidia releases a new GPU driver, I always look for info and benchmarks to find out if I should update my graphic driver or remain with the current one. I believe this is a problem afflicting many gamers as we want to achieve the best performance; but sometimes updating the drivers may cause issues and even reduce our fps.
The main idea behind this article is providing useful analyses and benchmarks on the performance of each driver with the main game engines. It is a work in collaboration with lokkenjp, who is the man behind the tests and comments displayed on this article.
This guide contains historical info and raw numbers about fps and frame times on all the Nvidia drivers from version 399.24 onwards.
Disclaimer: this is not a “professional” benchmark, but just some quick numbers and the subjective explanations of lokkenjp. You may have different results with your own PC depending on the hardware and software configuration, especially if you are using a card from a different generation/architecture than the one used for these benchmarks (Pascal 10xx GTX cards). Nevertheless, I believe this is a solid test to find out if it is worth upgrading driver version or not. Each driver benchmark compares the results with the previous one.
I have provided info on the PC used for the benchmarks, the games tested, the methodology and tools. If you prefer jumping directly to the tests, use the links below.
Quick links to each driver version
Benchmark PC
OS: Windows 10 64-bit (latest patches applied).
CPU: Intel i7-4790k. Stock clocks.
Memory: 16 GB DDR3-1600.
Graphic card: Asus Strix GTX 1070Ti Advanced Binned. Stock clocks.
Monitor: BenQ 1080p 60Hz. No G-Sync or HDR.
Notes:
For driver 416.34 and older versions the PC had 8GB of ram.
For driver 416.16 and older versions the PC had a different CPU: i5-4590.
Games tested
Tom Clancy’s: The Division 2
Note: game tested starting from driver 419.67.
Snowdrop Engine
DX12, 1080p resolution, High/Ultra settings (Volumetric Fog set to medium)
Ghost Recon: Wildlands
AnvilNext engine
DX11, 1080p resolution, mostly Very High settings but with Gameworks options disabled
FarCry 5
Dunia Engine (a heavily modified fork of the original CryEngine)
1080p resolution, maxed Ultra settings, TAA, FoV 90
Batman: Arkham Knight
Unreal Engine
1080p resolution, maxed settings, all Gameworks options enabled (thus, heavily using Nvidia PhysX engine)
Middle Earth: Shadow of Mordor
LithTech Engine
1080p resolution, maxed Ultra settings, FXAA
Tom Clancy’s: The Division
Note: game tested with driver 419.67 and older versions.
Snowdrop Engine
DX12, 1080p resolution, almost maxed settings (just lowered a bit Extra Streaming Distance and Object Detail), Neutral Lightning
Monster Hunter: World
Note: game tested from driver 417.71 to 418.91.
Capcom propietary MT Framework engine
Tested using the Kushala Daora engine-rendered cutscene
1080p resolution, maxed settings except for Volume Rendering and Subsurface Scattering which are both set to off, FXAA.
Methodology and tools
All the cinematic options (Motion Blur, Chromatic Aberration, Film Grain, Vignette effects, Depth of Field, etc.) are disabled when possible.
For driver 425.31 and newer versions
With previous drivers, I used the ingame FPS numbers, and FRAPS only for the lower Frame Times.
Now, I still use the same benchmark closed loops, but I use the FRAPS log for everything (FPS, average Frame Times, and lower Frame Times).
I’ve developed a custom tool that can parse several logs at once and does all the calculations with much more precision.
Only for The Division 2, we will continue using the numbers given by the game benchmark tool.
For driver 419.67 and older versions
Unless explicitly stated otherwise, all the games are run in borderless windowed mode, and tested using their own benchmarking tools.
430.64 is the Game Ready driver for Quake II RTX. It will available on June 6th, 2019.
430.xx version
430.86
I’m still on the v1809 version of Windows 10. The 1903 May update has been available to the public only for a few days, and I want to wait a couple of weeks so the initial issues can be ironed out by Microsoft before updating.
Having said that, for what I’ve been reading this latest Windows 10 v1903 update is much more stable than previous releases, and this will probably be the last test performed with the old October Update.
Tom Clancy’s: The Division 2
Three runs with 430.86:
Avg. FPS: 90.66 / 89.99 / 90.42
Frame times (3-run average): Avg. 11.07 - Lower 1% 13.04 - Lower 0.1% 17.13
The Division 2 under DX12 is performing exactly like the previous driver. Same average Frames Per Second, same lower Frame Times.
Ghost Recon: Wildlands
Three runs with 430.86:
Avg FPS: 79.74 / 78.78 / 78.99
Frame times (3-run average): Avg. 12.63 - Lower 1% 15.21 - Lower 0.1% 17.55
In the DX11 Wildlands we get again very similar results to 430.64. The lower 0.1% is a bit worse, but nothing to worry about, and it might be test noise, all within the error margin. Everything else is unchanged.
FarCry 5
Three runs with 430.86:
Avg FPS: 86.66 / 86.64 / 86.58
Frame times (3-run average): Avg. 11.54 - Lower 1% 15.48 - Lower 0.1% 18.34
While the average FPS on FarCry5 is more or less the same than the previous drivers, the lower frame times have been improved by quite a lot. Su much in fact that I had to retest just to be sure.
It is consistent. While the average performance is unchanged, the game is performing much smoother now, with fewer lag spikes and less stuttering all around. Good news at last!
Batman: Arkham Knight
Three runs with 430.86:
Avg FPS: 79.13 / 79.01 / 79.30
Frame times (3-run average): Avg. 12.64 - Lower 1% 20.69 - Lower 0.1% 26.05
Arkham Knight and its Unreal Engine is getting a very small but consistent hit in performance. About 1 less FPS on average, and equivalent losses on the lower Frame Times.
While we can consider anything below a 3% difference within the test error margin, the fact that the results are stable and consistent makes me believe it’s not test noise but a trend.
Middle Earth: Shadow of Mordor
Three runs with 430.86:
Avg FPS: 126.21 / 126.05 / 126.06
Frame times (3-run average): Avg. 7.93 - Lower 1% 11.21 - Lower 0.1% 13.40
Shadow of Mordor is more or less in line with the previous driver too. While the average FPS numbers seems a hair lower (again 1FPS down on average, mostly insignificant on a 120 refresh rate), the Frametimes are somewhat better, so the results are OK.
Driver testing results
So far the Driver itself is stable on my machine. Tested The Division 2, Wildlands, FarCry5, XCOM2, Anno 1800, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor, Endless Space 2, Diablo 3, StarCraft2, WoW and Magic The Gathering: Arena (short testing game sessions).
All ran fine without crashes or stability issues on my rig.
Performance is largely unchanged on all tests, except for a massive and unexpected improvement on the Lower Frame times of FarCry 5. I’m not aware of Ubisoft installing any update on FC5 recently, so I assume something was fixed within the drivers, either for FC5 or for FC: New Dawn (which shares same exact Dunia visual engine).
My recommendation
This one is pretty straightforward. If you are on Pascal and already updated to the 430 driver branch, this new driver seems safe, stable and brings some improvements to FarCry5. Plus the new Game Ready features, bug fixes, and new G-Sync compatible monitors.
While this testing is only valid under Windows 10 v1809, I assume the same might apply if you already updated to the new 1903 May Update.
For people staying on older versions of the drivers/OS, this new driver does not seem to bring anything new to the table performance-wise, and the only relevant changes are the already mentioned new Game Ready profiles or bug fixes, so if they don’t appeal to you, the performance alone of this new driver is not a reason to update.
World War Z performance with driver 430.64 - Nvidia
430.64
Note: For the record, I’m still on the old Windows 10 version, as the May update is only available for Preview Insiders at this time. So the optimizations implemented for WDDM 2.6 won’t be active on my tests.
Tom Clancy’s: The Division 2
Three runs with 430.64:
Avg. FPS: 90.51 / 90.10 / 90.26
Frame times (3-run average): Avg. 11.08 - Lower 1% 13.03 - Lower 0.1% 17.43
Same average Frames Per Second. The Lower 1% Frame time is better, while the Lower 0.1% is a hair worse (but within the margin of error). No big changes at all, although the improved Low 1% Frame Time is welcome (game seems more stable overall).
Ghost Recon: Wildlands
Three runs with 430.64:
Avg FPS: 79.79 / 78.82 / 78.72
Frame times (3-run average): Avg. 12.64 - Lower 1% 15.05 - Lower 0.1% 16.88
In the Dx11 Wildlands the bad results of the previous driver seems to have improved again. While the Average FPS are comparable to 430.39, the Lower Frame Times are quite better. which means better frame pacing, fewer stutters, and smoother gameplay. Good news so far!
FarCry 5
Three runs with 430.64:
Avg FPS: 85.95 / 88.36 / 86.41
Frame times (3-run average): Avg. 11.51 - Lower 1% 17.31 - Lower 0.1% 23.60
Average FPS/Frame Time is a bit better, Lower 1% Frame Time is a tiny bit worse, while the Lower 0.1% Frame Time is somewhat better. The changes are small enough to consider them within the test error margin.
Batman: Arkham Knight
Three runs with 430.64:
Avg FPS: 80.85 / 80.86 / 80.78
Frame times (3-run average): Avg. 12.37 - Lower 1% 20.38 - Lower 0.1% 25.37
More or less the same result of Far Cry 5. Some values are up and some down by a tiny amount. In the end, no significant changes at all.
Middle Earth: Shadow of Mordor
Three runs with 430.64:
Avg FPS: 127.41 / 127.06 / 127.23
Frame times (3-run average): Avg. 7.86 - Lower 1% 11.25 - Lower 0.1% 13.85
We have again the same picture with Middle Earth: Shadow of Mordor. The values are almost identical to 430.39, so we can observe no changes in this game.
Driver testing results
So far the Driver is stable on my machine. Tested The Division 2, Wildlands, FarCry5, XCOM2, Anno 1800, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor, Endless Space 2, Diablo 3, StarCraft2 and WoW (short testing game sessions).
After the relative disappointment of the first 430 driver release for Pascal 10xx cards, this one seems to be slightly better. The Division 2 seems to be running a bit smoother now with 430.64, and GR:Wildlands too. The other games are on par with 430.39
I’m still curious about the performance of this new driver branch under Windows 10 v1903, but we will have to keep waiting until Microsoft makes the new May Update available to the general public.
My recommendation
To be honest, I don’t know what to say about this new driver.
Performance-wise this one is almost the same or a tiny bit better than the previous driver. It does have some new G-Sync compatible (Freesync) monitors, a bunch of bug fixes, and three security vulnerability fixes.
Like the previous one, if you are running the Insider Preview of Windows 10 v1903, then probably you should install the new driver.
For people owning a Pascal card and still on the current Windows 10 v1809 (or previous OS versions), I still believe the 425 release is smoother overall. Unless you need one of the bug fixes or own one of the new G-Sync compatible monitors.
New G-Sync compatible monitors available with driver 430.39 - Nvidia
430.39
Note: Even if this new driver is WDDM 2.6 compatible (for the Windows 10 v1903 May Update release, now only available to Insider testers), I’m still testing the current Windows 10 v1809 version. This means that, even if this is a new driver branch with a lot of changes from the previous 425 driver, the driver will be still running on WDDM 2.5 and many optimizations won’t be fully functional until the new Win10 version is released.
Tom Clancy’s: The Division 2
Three runs with 430.39:
Avg. FPS: 90.58 / 90.22 / 90.18
Frame times (3-run average): Avg. 11.07 - Lower 1% 14.40 - Lower 0.1% 17.00
While the average performance is only slightly up - nothing significant - the Lower Frame times are better by a good amount. That means a more stable gameplay, with less game spikes and stuttering all around. It looks like The Division 2 under DX12 is definitely performing a bit better with this driver.
Ghost Recon: Wildlands
Three runs with 430.39:
Avg FPS: 79.29 / 78.48 / 78.49
Frame times (3-run average): Avg. 12.70 - Lower 1% 16.13 - Lower 0.1% 19.48
In the DX11 Wildlands game we find a different picture. The average framerate is about the same on both drivers, but the Lower Frame Times are a bit worse now, which means slightly worse stability with more stuttering.
FarCry 5
Three runs with 425.31 (before game 1.013 patch):
Avg FPS: 88.95 / 87.64 / 87.33
Frame times (3-run average): Avg. 11.37 - Lower 1% 15.19 - Lower 0.1% 18.55
Three runs with 425.31 (after game 1.013 patch):
Avg FPS: 85.42 / 86.58 / 85.12
Frame times (3-run average): Avg. 11.66 - Lower 1% 17.10 - Lower 0.1% 23.53
Three runs with 430.39 (after game 1.013 patch too):
Avg FPS: 86.97 / 85.42 / 84.16
Frame times (3-run average): Avg. 11.70 - Lower 1% 17.13 - Lower 0.1% 24.37
After some good news for FarCry 5 with the previous drivers, on April 18th I got a game update, which decreased the performance again by quite a chunk. It was so obvious that I had to retest the old 425 drivers with the updated FC5 to get new data points for the comparison. The tests confirmed that the game is performing quite worse (again, this is caused by a GAME update).
Testing the latest FarCry 5 version with the new drivers doesn’t show any improvement at all. In fact, all numbers are on a downward trend, even if by a very narrow margin.
Batman: Arkham Knight
Three runs with 430.39:
Avg FPS: 80.10 / 80.34 / 80.09
Frame times (3-run average): Avg. 12.47 - Lower 1% 20.53 - Lower 0.1% 25.10
Following the same trend of the other DX11 games, Arkham Knight is performing slightly worse than the previous driver.
Middle Earth: Shadow of Mordor
Three runs with 430.39:
Avg FPS: 127.44 / 127.40 / 127.52
Frame times (3-run average): Avg. 7.85 - Lower 1% 11.18 - Lower 0.1% 13.78
And closing the test is Middle Earth: Shadow of Mordor. The oldest game of the five tested is performing more or less like the 425 driver. No significant changes can be seen neither in the Average frame rate, nor in the Lower Frame Times.
Driver testing results
So far the Driver is stable on my machine. Tested The Division 2, Wildlands, FarCry5, XCOM2, Anno 1800, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor, Endless Space 2, Diablo 3, StarCraft2 and WoW (short testing game sessions).
All ran fine without crashes or stability issues on my rig.
Now with the interesting part. I had pretty big expectations for this new driver branch, as many people were reporting gains and improvements on their games with the 430.xx release. Unfortunately, even if DX12 on The Division 2 does seem to improve a bit, the Dx11 performance on all other games is equal or slightly worse than the previous driver (at least for Pascal/GTX 10xx cards)
Maybe the improvements reported could be traced to the new WDDM 2.6 features and optimizations instead than to the drivers themselves, so we will have to wait until Microsoft releases Win10 May Update (v1903) to the public. The official public release should be still a few weeks away in the best case, so maybe nVidia could release yet another driver revision meanwhile.
Anyway, I’ll test the new Windows 10 v1903 performance changes as soon as it becomes available to the general public.
My recommendation
This one is a bit tricky for me. If you are running the Insider preview of Windows 10 v1903, then probably you should install the new driver right away, as the previous driver betas for that OS had some nasty memory leaks.
Also, if you run DX12 games, and only DX12 games (like BF V, The Division 2,...) or Vulkan games, you might get some extra performance from this new release.
For everyone else playing DX11 games (or a mix of DX11 and DX12) with a Pascal GTX 10xx card, I think the new driver is (for now) not worth it. Slightly worse performance and increased lower frame times means worse frame pacing stability and more lag spikes.
This might change once the new Windows 10 v1903 is released into the wild.
For now, if you have a Pascal card and play DX11 games on Windows 10 v1809 or previous OS version, you should keep your current drivers unless you need some specific fix included on this version, or if you own any of the new FreeSync supported monitors.
Starting from driver 425.31, DXR is available also on non-RTX cards - Nvidia
425.xx version
425.31
Note: starting from this driver version, I have changed the methodology and the tools used for testing (more info on the “methodology and tools” section above). The results might not be directly comparable with the numbers given on previous driver benchmarks, but the data will be more accurate now. I have also tested again all games with 419.67 using the new tools.
Tom Clancy’s: The Division 2
Three runs with 419.67:
Avg. FPS: 89.46 / 90.13 / 90.05
Frame times (3-run average): Avg. 11.13 - Lower 1% 16.63 - Lower 0.1% 18.77
Three runs with 425.31:
Avg. FPS: 89.44 / 89.84 / 90.07
Frame times (3-run average): Avg. 11.14 - Lower 1% 16.47 - Lower 0.1% 18.07
So far the game is performing quite good for me on DX12. Performance between both drivers is almost identical, both on raw FPS and Frame Times. No changes at all here.
Ghost Recon: Wildlands
Three runs with 419.67:
Avg FPS: 78.18 / 79.26 / 78.20
Frame times (3-run average): Avg. 12.73 - Lower 1% 15.55 - Lower 0.1% 18.03
Three runs with 425.31:
Avg FPS: 79.43 / 78.62 / 78.61
Frame times (3-run average): Avg. 12.68 - Lower 1% 15.50 - Lower 0.1% 18.27
Wildlands is performing the same on both releases. This driver has a nice bug fix for Wildlands though, which caused some crashes in the game when opening the inventory menu.
FarCry 5
Three runs with 419.67:
Avg FPS: 83.92 / 86.61 / 84.59
Frame times (3-run average): Avg. 11.76 - Lower 1% 18.11 - Lower 0.1% 23.22
Three runs with 425.31:
Avg FPS: 88.95 / 87.64 / 87.33
Frame times (3-run average): Avg. 11.37 - Lower 1% 15.19 - Lower 0.1% 18.55
FarCry 5 got a pretty significant boost in this driver release. I don't know if it is thanks to some Far Cry: New Dawn optimizations, but whatever the reason the game is performing much better. Not only we have 2-3 more average FPS, but the lower Frame Times are substantially better, which means a smoother gameplay and less stuttering.
Batman: Arkham Knight
Three runs with 419.67:
Avg FPS: 81.45 / 80.87 / 81.86
Frame times (3-run average): Avg. 12.29 - Lower 1% 20.39 - Lower 0.1% 23.61
Three runs with 425.31:
Avg FPS: 80.87 / 80.33 / 81.04
Frame times (3-run average): Avg. 12.38 - Lower 1% 20.38 - Lower 0.1% 24.68
Arkham Knight is also stable on this driver. Averages might be a hair lower, but well within the error margin, and the Frame Times are more or less on the same values. Given that that 419.67 gave some small improvements to this game, the results in this one are good too.
Middle Earth: Shadow of Mordor
Three runs with 419.67:
Avg FPS: 127.88 / 127.79 / 127.70
Frame times (3-run average): Avg. 7.83 - Lower 1% 11.00 - Lower 0.1% 13.33
Three runs with 425.31:
Avg FPS: 127.41 / 127.26 / 127.44
Frame times (3-run average): Avg. 7.85 - Lower 1% 11.25 - Lower 0.1% 13.55
Like in the couple of previous tests, Middle Earth: Shadow of Mordor is another carbon copy of 419.67 results. Same raw FPS, same average Frame Time variance all around.
Driver testing results
System stability is so far good with the new driver. Tested The Division, The Division 2, Wildlands, FarCry5, XCOM2, Anno 2205, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, ME: Andromeda, Monster Hunter: World, Middle Earth: Shadow of Mordor, Endless Space 2 and WoW (short testing game sessions).
All ran fine without crashes or stability issues.
Regarding performance on Pascal cards (GTX 10xx), this driver have some very interesting improvements on FarCry 5. The rest of the tested games are more or less on the same FPS and Frame Time numbers as 419.67.
My recommendation
This one seems a no-brainer. Huge improvements on FarCry 5, DXR implementation (Raytracing compatibility) for non-RTX cards, some bug fixes, and so far the driver is pretty stable for me otherwise, with no glaring issues or performance regressions.
I think it's pretty safe to upgrade to this release for anyone already using the 4xx driver branch, (Pascal card users).
At this point I think that, even as 399.24 might be slightly better on old Dx11 games for 10xx cards, you should give this driver release a try. Remember that you might get that small bits of extra performance in exchange for lacking many game fixes and all patched security vulnerabilities disclosed on February on the 419.17 driver release (and more).
419.xx version
Driver 419.67
Tom Clancy’s: The Division
Three runs with 419.67:
Avg. FPS: 86.1 / 86.2 / 86.0
Typical FPS: 86.7 / 86.8 / 86.9
Avg. CPU: 66% / 59% / 61%
Avg. GPU: 96% / 96% / 96%
This will probably be the last test on The Division 1. Performance is largely unchanged. Perceived smoothness during gameplay is also the same as before. No lag spikes or stutters.
Tom Clancy’s: The Division 2
Three runs with 419.67:
Avg. FPS: 89.79 / 89.71 / 90.11
Avg. CPU: 81% / 79% / 70%
Avg. GPU: 97% / 97% / 97%
Frame times (3-run average): Avg. 11.13 - Lower 1% 14.75 - Lower 0.1% 15.
This is the first benchmark with The Division 2. So far the game is rock solid for me on DX12, and tons of fun too. Much more stable than TD1.
Three runs with 419.35:
Avg. FPS: 89.1 / 90.15 / 89.54
Avg. CPU: 79% / 76% / 78%
Avg. GPU: 97% / 97% / 97%
Frame times (3-run average): Avg. 11.16 - Lower 1% 14.72 - Lower 0.1% 15.58
As can be seen, performance between both drivers is almost identical, as are the Frame Times, which mean the same in-game smoothness.
Ghost Recon: Wildlands
Three runs with 419.67:
Avg FPS: 78.35 / 79.42 / 78.43
Min FPS: 69.10 / 67.66 / 68.59
Max FPS: 87.82 / 89.37 / 88.00
Avg CPU: 52.5% / 54.4% / 49.%
Avg GPU: 96.1% / 96.5% / 96.2%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.5 - Lower 0.1% 18.0
Another driver with flat changes on Wildlands. Raw Frames per Second numbers and Frame Times are again stable across releases.
FarCry 5
Three runs with 419.67:
Min FPS: 71 / 67 / 68
Avg FPS: 85 / 88 / 87
Max FPS: 106 / 109 / 107
Frame times (3-run average): Avg. 11.8 - Lower 1% 17.6 - Lower 0.1% 23.3
FarCry 5 seems to be working slightly worse on this new driver. While framerate is almost the same, maybe just a hair slower, the loss is more noticeable on the lower 1% Frame Times, which indicates a bit more stuttering during the benchmark.
Batman: Arkham Knight
Three runs with 419.67:
Min FPS: 41 / 42 / 42
Max FPS: 119 / 121 / 124
Avg FPS: 84 / 84 / 86
Frame times (3-run average): Avg. 12.3 - Lower 1% 20.4 - Lower 0.1% 23.7
Arkham Knight is performing slightly better on this driver. FPS and average Frame Times are the same, but the lower 0.1% Frame Time is lower (better) on a significant amount, so this driver seems to improve the smoothness and reduce lag spikes during playtime.
Middle Earth: Shadow of Mordor
Three runs with 419.67:
Avg FPS: 130.96 / 130.94 / 130.80
Max FPS: 188.99 / 193.02 / 189.66
Min FPS: 75.49 / 81.20 / 84.90
Frame times (3-run average): Avg. 7.82 - Lower 1% 11.0 - Lower 0.1% 13.3
Shadow of Mordor is another carbon copy of 419.35 results. Same raw FPS, same average Frame Time variance all around. No changes at all.
Driver testing results
System stability is so far good with the new driver. Tested The Division, The Division 2, Wildlands, FarCry5, XCOM2, Anno 2205, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, ME: Andromeda, Monster Hunter: World, Middle Earth: Shadow of Mordor, Endless Space 2 and WoW (short testing game sessions). All ran fine without crashes or stability issues.
Regarding Performance, this driver again seems more or less neutral for Pascal owners. While FarCry5 may be slightly worse on the Frame Time side, (maybe because of bug fixes to FarCry: New Dawn, which shares the same engine as FC5), Batman Arkham Knight on Unreal Engine seems a bit better.
My recommendation
Given the big amount of bug fixes included in this driver, new Game Ready optimizations for recent releases, the new Freesync profiles for two monitors, and more or less the same exact performance as the old one, I think it’s safe to upgrade to this release for anyone already on the 4xx driver branch.
As with the last few drivers, unless you happen to find a specific bug or performance regression on a specific game, this one seems a solid choice for those on the 4xx Driver Branch.
And also, the same caveat applies as previous drivers: if you are still on 399.24 (or older) that’s probably a better performer overall for DX11 games on Pascal 10xx cards. If you don’t have issues with your games, you might keep it.
Remember though that you are getting that extra performance in exchange of lacking the game fixes and all patched security vulnerabilities disclosed in February with the 419.17 driver release.
With driver 419.67, DLSS technology is now available in Anthem, with up to 40% improvement on gameplay performance according to Nvidia.
Driver 419.35
Tom Clancy’s: The Division
Three runs with 419.35:
Avg. FPS: 86.3 / 86.3 / 86.4
Typical FPS: 86.8 / 86.8 / 86.9
Avg. CPU: 66% / 64% / 63%
Avg. GPU: 96% / 96% / 96%
DX12 performance on The Division using Pascal is largely unchanged. Perceived smoothness during gameplay is also identical to 419.17. No lag spikes or stutters.
Ghost Recon: Wildlands
Three runs with 419.35:
Avg FPS: 79.79 / 78.50 / 78.54
Min FPS: 69.79 / 69.24 / 69.65
Max FPS: 89.91 / 88.00 / 88.12
Avg CPU: 51.6% / 50.9% / 56.2%
Avg GPU: 96.5% / 96.2% / 96.4%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.9 - Lower 0.1% 18.3
Once again, the DX11 Wildlands benchmark is also almost equal to the last driver both in raw FPS numbers and on Frame Times as well.
FarCry 5
Three runs with 419.35:
Min FPS: 71 / 73 / 69
Avg FPS: 86 / 89 / 88
Max FPS: 108 / 108 / 109
Frame times (3-run average): Avg. 11.5 - Lower 1% 15.9 - Lower 0.1% 22.8
FarCry 5 is still performing good. No significant changes on Frames per Second.
Frame times are a bit of a mixed bag. Average is the same, the lower 1% is quite a bit better, while the lower 0.1% is a bit worse. That could mean better stability overall, but with a punctual hiccup or two scattered here and there.
Batman: Arkham Knight
Three runs with 419.35:
Min FPS: 43 / 43 / 43
Max FPS: 123 / 120 / 122
Avg FPS: 84 / 84 / 85
Frame times (3-run average): Avg. 12.3 - Lower 1% 20.5 - Lower 0.1% 25.6
Arkham Knight is performing slightly better on this driver. While FPS and the average frame times are about the same, both the lower 1% and 0.1% Frame Times are lower (better) by a non-trivial amount.
Middle Earth: Shadow of Mordor
Three runs with 419.35:
Avg FPS: 131.68 / 131.42 / 130.93
Max FPS: 191.76 / 192.55 / 190.87
Min FPS: 73.87 / 81.55 / 86.13
Frame times (3-run average): Avg. 7.81 - Lower 1% 11.0 - Lower 0.1% 13.4
Shadow of Mordor behavior is exactly the same as it was on 419.17. Same raw Frames per Second, same average Frame Time variance.
Driver testing results
System stability is so far good with the new driver. Tested The Division, Wildlands, FarCry5, XCOM2, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, ME: Andromeda, Monster Hunter: World, Middle Earth: Shadow of Mordor, Star Trek Online, Endless Space 2 and WoW (short testing game sessions). All ran fine without crashes or stability issues.
Speaking about Performance, this driver seems neutral for Pascal owners. Even as I recorded some improvements outside margin of error on Batman Arkham Knight frame times, this result seems isolated and maybe it could be somehow linked to the fix included in this driver for PhysX fog in the Arkham Origins game, (or maybe just it’s an outlier result).
Save for that game, the rest is on “flat line”.
New driver have nevertheless a very interesting fix for Apex Legends, and also Game Ready optimizations for this same title. Also, it includes another fix for high CPU usage of the NVDisplay container, another Hitman 2 graphical glitch fixed, and The Division 2 optimized profile (which I’m eager to test on the live version as soon as it releases on March 15th).
My recommendation
This one seems a no-brainer. If you installed the previous 418/419 drivers on your Pascal cards, there is no reason not to upgrade to this latest 419.35. Same or better performance on the tested games, and fixes for several games (including Apex Legends, the current “king of the hill” of survival games).
Unless you happen to find a specific bug or performance regression on a specific game, this one seems a solid choice for those already on the 4xx Driver Branch.
Like happened on previous drivers, if you are still on 399.24 (or older) that’s probably a better performer overall for DX11 games on Pascal 10xx cards, and if you don’t have issues you might keep it. Remember though that you are getting that extra performance in exchange of lacking the fixes for all vulnerabilities disclosed on the previous driver release, along with the unpatched bugs on recent games of this one.
419.35 is the Game Ready driver for Apex Legends - Nvidia
Driver 419.17
Note: Windows 10 v1809 have just received a Patch which enabled the “Retpoline” solution to Spectre v2 mitigation, which hopefully will improve the overall performance of systems with CPU bottlenecks and the Meltdown/Spectre patches enabled.
I’ve tested this new Win10 update both with the previous 418.91 driver and with the new 419.17 one, and I’ve not found any significant difference at all on any of my tested games.
Tom Clancy’s: The Division
Three runs with 419.17:
Avg. FPS: 86.6 / 86.4 / 86.8
Typical FPS: 87.1 / 86.8 / 87.2
Avg. CPU: 64% / 60% / 61%
Avg. GPU: 96% / 96% / 96%
DX12 performance on Pascal seems identical to previous driver.
Perceived smoothness during actual gameplay is also indistinguishable from 418. No lag spikes or stutters.
Ghost Recon: Wildlands
Three runs with 419.17:
Avg FPS: 79.29 / 78.52 / 78.18
Min FPS: 69.31 / 69.17 / 68.59
Max FPS: 90.19 / 88.91 / 88.56
Avg CPU: 52.2% / 51.8% / 52.5%
Avg GPU: 96.6% / 96.4% / 96.1%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.8 - Lower 0.1% 18.7
Like The Division, the DX11 Wildlands benchmark is also almost equal both in raw FPS numbers and on Frame Times as the previous driver.
FarCry 5
Three runs with 419.17:
Min FPS: 70 / 67 / 68
Avg FPS: 88 / 86 / 88
Max FPS: 108 / 107 / 108
Frame times (3-run average): Avg. 11.6 - Lower 1% 16.8 - Lower 0.1% 21.6
FarCry 5 stays strong on this new driver. Again, no significant changes neither with Raw FPS nor with the Frame Times. The lower 0.1% is a tad better, but may very well be within the error margin.
Batman: Arkham Knight
Three runs with 419.17:
Min FPS: 42 / 41 / 42
Max FPS: 121 / 125 / 124
Avg FPS: 84 / 84 / 84
Frame times (3-run average): Avg. 12.3 - Lower 1% 21.7 - Lower 0.1% 27.2
Arkham Knight follows the same trend as the other games. Frame times are maybe a tiny bit worse, but nothing to worry about.
Middle Earth: Shadow of Mordor
Three runs with 419.17:
Avg FPS: 131.12 / 131.01 / 130.90
Max FPS: 191.96 / 190.84 / 189.76
Min FPS: 79.48 / 85.10 / 85.38
Frame times (3-run average): Avg. 7.83 - Lower 1% 11.0 - Lower 0.1% 13.0
Shadow of Mordor is also almost a carbon copy of the 418.91 results.
Driver testing results
System stability is so far good with the new driver. The Division, Wildlands, FarCry5, XCOM2, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, ME: Andromeda, Monster Hunter: World, Middle Earth: Shadow of Mordor, Star Trek Online, Endless Space 2 and WoW (short testing game sessions) all ran fine without crashes or stability issues.
I must say too that The Division 2 Open Beta is running great with this new drivers (which hopefully will replace TD1 as the DX12 benchmark as soon as it’s released on March 15th).
I’ve got some strange readings while testing Monster Hunter World, and given that the game does not have a built-in benchmark tool, I decided to pull it off. Performance on this game is good though (I’d say, at least equal to 418.91)
Performance-wise this driver is a rather dull one. On my Pascal machine I’m reading about the same exact performance values as the previous drivers on my tested games. Which, on the other hand, is good enough given that this driver release is including a big bunch of security fixes.
My recommendation
Given the big list of security fixes included and the performance stability, I’m recommending this driver for any Pascal user which already jumped to the 4xx driver branch.
Like happened on previous drivers, if you are still on 399.24 that’s probably a better performer overall for DX11 games on Pascal 10xx cards, and if you don’t have issues you might very well keep it. Have in mind though that you are getting that extra performance in exchange of having some nasty vulnerabilities unpatched on your drivers.
419.67 introduced support for Nvidia Surround when using G-SYNC Compatible displays - Nvidia
418.xx version
Driver 418.91
Tom Clancy’s: The Division
Three runs with 418.91:
Avg. FPS: 86.8 / 86.4 / 86.7
Typical FPS: 87.3 / 86.9 / 87.1
Avg. CPU: 64% / 58% / 61%
Avg. GPU: 96% / 96% / 96%
DX12 performance on Pascal remains stable. All numbers stay almost on par with 418.81
Perceived smoothness during actual gameplay feels in line with the previous driver. No lag spikes or stutters.
Ghost Recon: Wildlands
Three runs with 418.91:
Avg FPS: 79.67 / 78.95 / 79.11
Min FPS: 68.11 / 69.93 / 68.93
Max FPS: 90.73 / 88.91 / 89.11
Avg CPU: 52.0% / 51.6% / 53.8%
Avg GPU: 96.4% / 96.7% / 96.8%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.5 - Lower 0.1% 17.9
DX11 Wildlands benchmark is another carbon copy of the previous driver. Some numbers above, some below, all within a narrow error margin. Frame Times got a bit better on the last driver for me, and they are holding strong in this release. No smoothness changes.
FarCry 5
Three runs with 418.91:
Min FPS: 70 / 72 / 67
Avg FPS: 86 / 88 / 88
Max FPS: 107 / 109 / 108
Frame times (3-run average): Avg. 11.5 - Lower 1% 16.7 - Lower 0.1% 22.2
FarCry 5 improved quite a lot on 418.81, and is still performing great with the new driver. Averages are even a hair better (with better Frame time average), while the Lower Frame Time 1% also improves a bit.
Batman: Arkham Knight
Three runs with 418.91:
Min FPS: 40 / 39 / 41
Max FPS: 122 / 123 / 125
Avg FPS: 85 / 84 / 85
Frame times (3-run average): Avg. 12.2 - Lower 1% 21.2 - Lower 0.1% 26.9
Once again, all numbers stay in the same range as the previous driver. No noticeable changes at all.
Middle Earth: Shadow of Mordor
Three runs with 418.91:
Avg FPS: 131.71 / 131.33 / 131.90
Max FPS: 193.02 / 191.66 / 190.78
Min FPS: 78.29 / 84.33 / 84.02
Frame times (3-run average): Avg. 7.80 - Lower 1% 11.1 - Lower 0.1% 13.3
Shadow of Mordor follows the same trend as all other games. Almost the same raw performance as 418.81, maybe a hair better on minimum values, while also equivalent (or a tiny bit better) Frame Time values.
Changes are so small though that may be within the error margin.
Monster Hunter: World
Four runs with 418.91:
Avg FPS: 117 / 116 / 116
Lower 1% FPS: 81 / 83 / 84
Frame times (4-run average): Avg. 8.61 - Lower 1% 12.0 - Lower 0.1% 13.1
Monster Hunter World is yet again performing almost identically to previous driver.
Driver testing results
System stability is so far good with the new driver. The Division, Wildlands, FarCry5, XCOM2, Terraria, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor, Star Trek Online, Endless Space 2 and WoW (short testing game sessions) all ran fine without crashes or stability issues.
Performance-wise this driver does not have any noticeable change for Pascal cards. Given the small number version jump, and the focus on DLSS and RTX features (only for Turing users), this is not unexpected.
My recommendation
If you are still on 399.24, that’s probably still a better performer for “old” DX11 games on 10XX cards, and if you don’t have issues you might very well keep it.
But if you have already updated to the 4xx branch, this driver is pretty solid. Anyone who already installed 418.81 should update for the bugfixes, as performance is pretty much unchanged.
Real-time Ray Tracing in Metro Exodus. 418.91 provides the optimal gaming experience in this game when using DLSS and Ray Tracing.
Driver 418.81
Tom Clancy’s: The Division
Three runs with 418.81:
Avg. FPS: 86.6 / 86.3 / 86.5
Typical FPS: 87.1 / 86.8 / 86.9
Avg. CPU: 67% / 62% / 63%
Avg. GPU: 96% / 96% / 96%
Once again, seems DX12 performance is stable or even a tiny bit better than the previous driver. It’s not by much, and may be within the error margin of the test, but the trend seems consistent.
Actual gameplay feels in line with the previous driver. Smooth and stable.
Ghost Recon: Wildlands
Three runs with 418.81:
Avg FPS: 79.46 / 78.75 / 78.55
Min FPS: 67.06 / 69.31 / 67.66
Max FPS: 90.10 / 89.29 / 89.46
Avg CPU: 50.2% / 51.0% / 49.0%
Avg GPU: 96.2% / 96.4% / 96.5%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.4 - Lower 0.1% 17.4
This time, the DX11 Wildlands benchmark is better and more stable than the previous driver. Raw FPS numbers are a tiny bit better, nothing stellar, but what is more important, the Lower 1% and Lower 0.1% Frame Times are noticeably better, indicating an improvement on game stuttering and smoother gameplay.
Actual gameplay outside the benchmark run seems also smoother during fast action sequences.
FarCry 5
Three runs with 418.81:
Min FPS: 68 / 71 / 67
Avg FPS: 84 / 87 / 85
Max FPS: 108 / 108 / 108
Frame times (3-run average): Avg. 12.0 - Lower 1% 17.5 - Lower 0.1% 22.1
FarCry 5 have got a huge performance improvement in this driver. Whatever happened with 417.35 seems mostly fixed (lost about 10% performance back on that driver, for what I heard related to a DirectX update done by Microsoft on a Win10 update).
Whatever the cause, the performance may not be as good as it was on 417.22, but it is much better now than all other drivers in between.
Not only FPS averages are better, but also all Frame Time values are noticeably better, indicating a smoother gameplay with less stuttering.
Batman: Arkham Knight
Three runs with 418.81:
Min FPS: 41 / 42 / 41
Max FPS: 121 / 120 / 123
Avg FPS: 85 / 85 / 85
Frame times (3-run average): Avg. 12.3 - Lower 1% 21.3 - Lower 0.1% 26.3
While the Max FPS numbers are a bit down, the Average and Minimum values are up by a couple of extra FPS.
Not only that, but following the same trend the average Frame Time variance is somewhat better all around, and the Lower 0.1% improved by quite a chunk, indicating less lag spikes during the game.
Middle Earth: Shadow of Mordor
Three runs with 418.81:
Avg FPS: 131.87 / 131.24 / 131.23
Max FPS: 189.68 / 191.90 / 195.21
Min FPS: 77.30 / 77.02 / 79.87
Frame times (3-run average): Avg. 7.82 - Lower 1% 11.2 - Lower 0.1% 13.6
The huge changes which happened on Shadow of Mordor on 417.71 reverted back. This time the average FPS and maximum FPS are back to 417.22 numbers (lower than the previous driver), but the minimum FPS numbers are much better. At the same time, the Frame Times are substantially better, specially on the lower 1% and 0.1% brackets.
While the raw FPS average may be a bit lower, the game smoothness is now miles ahead of 417.71, with much less stuttering, and to be honest, not a single lag spike can be perceived during gameplay at all.
Monster Hunter: World
Three runs with 418.81:
Avg FPS: 116 / 116 / 115
Lower 1% FPS: 92 / 94 / 93
Frame times (4-run average): Avg. 8.68 - Lower 1% 10.7 - Lower 0.1% 12.4
Monster Hunter World shows a small decrease in the average FPS numbers (and the average Frame Time), yet the lower 1% FPS and Frame Times are a bit better. Thus, game is performing a bit slower overall, but more smooth and with better frame pacing, improving the gaming perception during both the benchmarked run and during actual gameplay (I hunted a Xeno’jiva and the game went butter smooth during the whole combat, not a single hiccup).
Driver testing results
System stability is so far good with the new driver. The Division, Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor, Star Trek Online and WoW (short testing game sessions) all ran fine without stability issues.
Performance-wise this driver is good for Pascal cards. Really good. In fact, given the recent trend, even surprisingly good. Not only all games got improvements in key areas (smoothness and even Raw FPS averages on most tests), but the massive drop in performance of FarCry 5 on 417.35 seems to have been corrected.
Performance seems a bit below 417.22 still on DX11 games, yet I’d say that it’s better overall than any other 4xx branch driver release (except for the mentioned 417.22).
My recommendation
This time, I have no issue recommending this driver for Pascal owners which jumped into the 4xx driver branch. If you are still on 399.24, that’s probably still a better performer for DX11 games, and if you don’t have issues, you should keep it.
But if you need to update your drivers past 399 due to any bug, or if you are interested on DX12 performance improvements, Freesync compatibility, or any of the recent Game Ready profiles, then this new driver is probably the better choice (even more so for Laptop users who couldn’t run 417.22 due to the low power state bug).
A note here though. I’ve read from a couple of different sources of very poor performance of this driver on Total War Warhammer 2 (and maybe other Total War titles). I cannot confirm it, and I haven’t observed any drop in my tested games. But it may be a Total War engine specific issue, so have this in mind if you decide to update.
RTX GPUs are up to 15% more efficient than Pascal ones when streaming in H.264 - Nvidia
417.xx version
Driver 417.71
Tom Clancy’s: The Division
Three consecutive runs with 417.71:
Avg. FPS: 85.9 / 86.1 / 86.0
Typical FPS: 86.3 / 86.7 / 86.5
Avg. CPU: 67% / 63% / 62%
Avg. GPU: 96% / 96% / 96%
This new driver seems to increase performance a tiny bit on The Division. Numbers are about 1 FPS higher all around. It’s not much, and may be within the error margin of the test, but the trend seems consistent.
Actual gameplay feels like the previous driver, or even a hair smoother.
Ghost Recon: Wildlands
Three consecutive runs with 417.71:
Avg FPS: 78.03 / 77.54 / 77.95
Min FPS: 66.87 / 70.58 / 70.58
Max FPS: 89.91 / 87.74 / 87.74
Avg CPU: 52.1% / 51.9% / 51.4%
Avg GPU: 95.8% / 95.1% / 95.7%
Frame times (3-run average): Avg. 12.8 - Lower 1% 16.8 - Lower 0.1% 20.1
On the other hand, the gains on DX12 The Division seems to be lost on the DX11 Wildlands test. We can see a small but consistent decrease on FPS. It is significant that now the game is taking more average CPU percentage, while using fewer GPU resources, which could mean that this new driver is taxing the CPU a little more than the previous one for whatever reason.
Frame time pacing follows the same trend as the raw average FPS numbers. The average is just slightly worse, but the lower 1% and 0.1% frame times are noticeably worse, indicating an increase on game stuttering and less smooth gameplay.
FarCry 5
Three consecutive runs with 417.71:
Min FPS: 63 / 63 / 59
Avg FPS: 78 / 80 / 79
Max FPS: 104 / 103 / 102
Frame times (3-run average): Avg. 12.9 - Lower 1% 20.6 - Lower 0.1% 24.5
The huge performance hit we’ve got in the 417.35 driver (lost about 10-15% performance when compared to 417.22) is still here. For what I’ve read, that could be the result of some kind of weird interaction between the Nvidia drivers and some DirectX updates that Microsoft pushed on December Patch on Tuesday. Whatever the root cause could be, reverting to 417.22 solved the issue, while both 417.35 and 417.71 suffer a major hit.
While the FPS averages and the frame time averages have recovered a tiny bit, both the lower 1% and 0.1% frame times are noticeably worse again, indicating once more an increase of stuttering ingame and less smooth gameplay.
Batman: Arkham Knight
Three consecutive runs with 417.71:
Min FPS: 41 / 53 / 43
Max FPS: 122 / 123 / 120
Avg FPS: 82 / 84 / 82
Frame times (3-run average): Avg. 12.4 - Lower 1% 21.4 - Lower 0.1% 27.5
Average FPS performance is once again a bit down on Batman Arkham Knight.
Surprisingly, while the average Frame Time variance and the Lower 1% are somewhat lower, the Lower 0.1% is in fact a bit better, which may indicate a lower overall performance, but fewer stutters and lag spikes.
Middle Earth: Shadow of Mordor
Three consecutive runs with 417.71:
Avg FPS: 135.17 / 134.74 / 134.65
Max FPS: 232.05 / 234.95 / 231.00
Min FPS: 52.35 / 54.06 / 52.48
Frame times (3-run average): Avg. 7.90 - Lower 1% 16.3 - Lower 0.1% 18.9
Shadow of Mordor have suffered massive changes. While the average FPS and maximum FPS are higher, the lower FPS and the frame times are noticeably worse.
Frame time numbers recorded are extremely worse than previous driver (consistent with the wild variance on Min/Max FPS). This again shows worse game stability and more stuttering during the game.
Monster Hunter: World
Finally, a surprise for Monster Hunter World players. I’m including it from now on.
Four consecutive runs with 417.71:
Avg FPS: 120 / 120 / 119 / 119
Lower 1% FPS: 89 / 93 / 94 / 92
Frame times (4-run average): Avg. 8.39 - Lower 1% 10.8 - Lower 0.1% 12.5
Four consecutive runs with 417.35:
Avg FPS: 119 / 119 / 118 / 118
Lower 1% FPS: 89 / 93 / 92 / 91
Frame times (4-run average): Avg. 8.42 - Lower 1% 10.9 - Lower 0.1% 12.7
Monster Hunter World show a steady trend with this driver. Almost the same performance, almost the same Frame time numbers.
Driver testing results
System stability is so far good with the new driver. The Division, Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor and WoW (short testing game sessions) all ran fine without stability issues.
Performance wise this driver is once again worse than the previous one for Pascal users, at least on DX11 games. Except for a small performance increase on The Division (maybe because it’s the only DirectX 12 game), and MH:W which is stable, the rest of the tested games have worse average FPS numbers, worse frame pacing stability, or both.
The huge performance regression observed on FarCry 5 in the previous driver release is still here too. Whatever the underlying reason may be (Microsoft DirectX updates or whatever), the fact is that rolling back to 417.22 increases back the performance by >10%.
Certainly the 4xx driver branch, except for a couple of exceptions, is so far very disappointing for Pascal card owners.
My recommendation
In the end, I cannot recommend this driver at all for Pascal card owners (at least while playing DX11 games - the vast majority right now). The most relevant new features are support for RTX 2060 (which is pretty irrelevant for us) and working G-Sync with Freesync monitors. This one is a welcome change, and maybe the only reason why a Pascal owner might be interested in updating.
For the rest of us, I think it’s safe to say that this driver is not worth it. 399.24 is maybe the best overall recent driver, and 417.22 maybe the only 4xx branch driver worth updating, unless you are using a laptop card.
Driver 417.71 introduced G-Sync compatible monitors - Image: Nvidia - Check out our community list with info on hundreds of Freesync monitors.
Driver 417.35
Tom Clancy’s: The Division
Three consecutive runs with 417.35:
Avg. FPS: 85.1 / 85.0 / 85.1
Typical FPS: 85.7 / 85.6 / 85.6
Avg. CPU: 64% / 65% / 62%
Avg. GPU: 96% / 96% / 96%
The driver shows a very small but consistent 1 FPS decrease on all testing runs. Subjective smoothness perception and stuttering during benchmark and during regular gameplay is also a bit worse. I’ve got a few stutters here and there that just wasn’t there with previous drivers.
Ghost Recon: Wildlands
Three consecutive runs with 417.35:
Avg FPS: 79.53 / 78.71 / 78.94
Min FPS: 67.66 / 68.79 / 68.89
Max FPS: 90.10 / 88.29 / 89.26
Avg CPU: 49.4% / 47.6% / 49.2%
Avg GPU: 96.5% / 96.5% / 96.4%
Frame times (3-run average): Avg. 12.6 - Lower 1% 15.4 - Lower 0.1% 18.9
Wildlands have about the same performance all around. No visible changes.
Nevertheless, we have a significant worse lower 0.1% frame time, which mean increased game stuttering and worse frame pacing stability.
FarCry 5
Three consecutive runs with 417.35:
Min FPS: 60 / 61 / 60
Avg FPS: 78 / 78 / 79
Max FPS: 99 / 100 / 100
Frame times (3-run average): Avg. 13.1 - Lower 1% 17.5 - Lower 0.1% 21.5
We have a HUGE performance regression here. Almost 10 FPS less on all metrics (which is about a 15% loss in performance).
Frame times are consistent with this, and all three numbers are way worse than the previous driver.
Something has definitely happened there. I’ve double checked my numbers and even run some additional tests. The results are consistent. Game runs about 15% worse with this driver.
Finally, uninstalling with DDU 417.35 and going back 417.22 instantly returned numbers to previous levels, so it’s not an environmental factor, the driver is to blame for sure.
Batman: Arkham Knight
Three consecutive runs with 417.35:
Min FPS: 39 / 43 / 41
Max FPS: 125 / 125 / 124
Avg FPS: 83 / 84 / 84
Frame times (3-run average): Avg. 12.3 - Lower 1% 20.9 - Lower 0.1% 29.6
Again, we have a small performance loss, about 2 FPS less on average numbers.
Not only this, but also Frame Time variance is also somewhat worse with the new driver on all three metrics, specially on the Lower 0.1% (which means, again, worse frame pacing, more stuttering, and less smoothness during the game).
Middle Earth: Shadow of Mordor
Three consecutive runs with 417.35:
Avg FPS: 131.67 / 131.02 / 130.89
Max FPS: 194.69 / 192.30 / 191.81
Min FPS: 61.60 / 78.39 / 71.14
Frame times (3-run average): Avg. 7.82 - Lower 1% 11.4 - Lower 0.1% 13.5
Shadow of Mordor follows the same trend as the other games. While average FPS are comparable to the previous driver, minimum FPS numbers are noticeably worse.
Frame time numbers recorded are again somewhat worse than 417.22, resulting in a less smooth playing experience with increased lag spikes and stuttering.
Driver testing conclusions
System stability is so far good with the new driver. The Division (except for a reproducible crash to desktop during benchmark which I already know how to bypass), Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor and WoW (short testing game sessions) all ran fine without stability issues.
Performance wise this driver seems a bit worse than the last one, at least for Pascal cards owners. Quite worse in fact under certain games.
While the fix to the Mobile cards is welcomed for those affected (it was a pretty serious issue), the performance loss is noticeable in all games tested, with worse Frame Pacing stability all around, and it’s specially bad on Farcry 5. Almost 15% of the framerate is lost due to this driver. That’s an unbelievably big loss, which just does not have any meaningful justification.
My recommendation
In the end, I cannot recommend upgrading to this driver to anyone with a Pascal card which can run the previous 417.22. If you have a 10xx series card and you are not affected by the laptop issue locking the performance mode of the Nvidia card to lower power levels, better stay away from this driver and install the previous 417.22, which was a pretty good one.
417.35 enabled DLSS (Beta support) on Final Fantasy XV.
Driver 417.22
Tom Clancy’s: The Division
Three consecutive runs with 417.22:
Avg. FPS: 85.9 / 86.0 / 86.1
Typical FPS: 86.4 / 86.3 / 86.5
Avg. CPU: 61% / 65% / 63%
Avg. GPU: 96% / 96% / 96%
The driver shows almost identical FPS performance to previous 417.01 driver. No changes at all.
The subjective smoothness perception during the benchmark is maybe a hair better though. Also during gameplay (outside the benchmark loop), I’ve got less laggy spikes all around.
I’ve noticed too a small but consistent decrease in CPU usage during benchmarks (which, without any loss in performance, is good).
Ghost Recon: Wildlands
Three consecutive runs with 417.22:
Avg FPS: 80.04 / 79.19 / 78.96
Min FPS: 68.79 / 69.19 / 68.59
Max FPS: 90.46 / 89.29 / 89.11
Avg CPU: 50.3% / 47.2% / 46.5%
Avg GPU: 96.6% / 96.9% / 96.9%
Frame times (3-run average): Avg. 12.6 - Lower 1% 15.2 - Lower 0.1% 17.8
Wildlands have slightly higher FPS numbers, nothing relevant tbh, but with another noticeable decrease in CPU usage. Also, while the average Frametime is almost equal to 417.01, the lower 1% and 0.1% are significantly better, which means increased game smoothness and stability, with less stuttering.
FarCry 5
Three consecutive runs with 417.22:
Min FPS: 69 / 71 / 70
Avg FPS: 87 / 89 / 88
Max FPS: 108 / 110 / 110
Frame times (3-run average): Avg. 11.5 - Lower 1% 15.6 - Lower 0.1% 19.3
Same results as Wildlands. FPS numbers are a carbon copy of the previous results, but once again the Lower 1% and the Lower 0.1% frame times are better than 417.01.
Framerate seems more stable and with less stuttering all around.
Batman: Arkham Knight
Three consecutive runs with 417.22:
Min FPS: 40 / 40 / 39
Max FPS: 123 / 124 / 124
Avg FPS: 86 / 85 / 86
Frame times (3-run average): Avg. 12.2 - Lower 1% 20.6 - Lower 0.1% 26.2
We don’t have any significant change on Arkham Knight. FPS numbers are identical to previous driver.
Frame time numbers are also consistent with previous driver. Maybe a small regression, but changes are so small that may very well be test variance.
Middle Earth: Shadow of Mordor
Three consecutive runs with 417.22:
Avg FPS: 131.18 / 131.22 / 131.37
Max FPS: 191.88 / 192.72 / 191.41
Min FPS: 81.59 / 86.29 / 80.61
Frame times (3-run average): Avg. 7.81 - Lower 1% 11.1 - Lower 0.1% 12.9
Shadow of Mordor is tied in performance with the previous driver. FPS are on the same range, and Frame time numbers recorded are equal on Average and lower 1%, but have some improvement on the lower 0.1%.
Driver testing conclusions
System stability is good with the new driver. The Division (except for a reproducible crash to desktop during benchmark which I already know how to bypass), Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor and WoW (short testing game sessions) all ran fine without stability issues.
I must say though that several reports show a significant performance loss on laptops running Pascal cards (and maybe others) with this new drivers. It seems there is some kind of issue on certain laptop configurations with the new drivers and the graphics card Core Clock speed, which is being locked for some unknown reason to a maximum of 700Mhz. This in turn tanks performance badly as expected. Rolling back to 417.01 seems to fix the problem.
For desktop card owners, at least under Pascal architecture this driver seems to work pretty well. While average Frames per Second do not change, overall frame pacing stability and smoothness seems to improve, with less stutters, while also decreasing CPU usage all around.
My recommendation
In the end, I recommend upgrading to this driver to anyone with a Desktop card and already in the 4xx.xx Nvidia driver branch. Not only there are some bugfixes and new game Ready Profiles, but seems game smoothness is going in the good direction again. And maybe, even those still on the excellent 399.24 can even give this one a shot (you can always roll back to 399 using DDU if you find issues on your computer).
For Nvidia laptop owners it’s a very different story, though. There are some nasty reports of cards being locked in low power states, which it turn lower performance to unacceptable levels. Issue doesn’t seem to be universal, but spread enough to deserve a warning at least. If you are a laptop user, be sure to have DDU ready to roll back to 417.01 if you want to try the newest release.
417.22 is the Game Ready driver for Insurgency Sandstorm - Nvidia
Driver 417.01
Tom Clancy’s: The Division
Three consecutive runs with 417.01:
Avg. FPS: 86.1 / 86.1 / 86.2
Typical FPS: 86.6 / 86.7 / 86.8
Avg. CPU: 67% / 65% / 66%
Avg. GPU: 96% / 96% / 96%
Again, this driver show identical performance to previous 416.94 drivers, as is the subjective smoothness perception during benchmark. No changes at all.
Ghost Recon: Wildlands
Three consecutive runs with 417.01:
Avg FPS: 79.77 / 78.85 / 78.69
Min FPS: 68.48 / 69.93 / 68.79
Max FPS: 90.82 / 89.11 / 88.56
Avg CPU: 53.4% / 53.5% / 54.7%
Avg GPU: 96.3% / 96.2% / 96.4%
Frame times (3-run average): Avg. 12.7 - Lower 1% 16.4 - Lower 0.1% 20.3
Wildlands is also almost on par with previous 416.94. This time averages are a hair lower, but well within test variance limits.
Subjective game smoothness is also identical, no noticeable stutters at all.
Frame time numbers are also almost the same; average is identical, lower 1% is slightly worse, while at the same time lower 0.1% is a hair better.
FarCry 5
Three consecutive runs with 417.01:
Min FPS: 66 / 71 / 69
Avg FPS: 87 / 89 / 88
Max FPS: 109 / 109 / 110
Frame times (3-run average): Avg. 11.5 - Lower 1% 16.4 - Lower 0.1% 21.5
FPS results are this time a bit lower than the previous driver release. Averages are about 1 FPS down, minimum FPS a couple of units down or so.
What it’s interesting though is that while the Average frame time is a hair worse (consistent with the lower average FPS measure), both the Lower 1% and the Lower 0.1% are indeed a bit better than the previous driver.
As such, this game have lower Frames per Second on average, but framerate is more stable and with less stuttering instead.
Batman: Arkham Knight
Three consecutive runs with 417.01:
Min FPS: 42 / 39 / 40
Max FPS: 125 / 123 / 124
Avg FPS: 85 / 84 / 84
Frame times (3-run average): Avg. 12.1 - Lower 1% 20.2 - Lower 0.1% 25.5
We don’t have any significant change on Arkham Knight. All FPS measures are almost on par with previous driver.
Frame time numbers also show a slight improvement on the Lower 1% and Lower 0.1%, which should be understood as better frame pacing and stability.
Middle Earth: Shadow of Mordor
Three consecutive runs with 417.01:
Avg FPS: 131.52 / 131.21 / 131.32
Max FPS: 198.28 / 191.25 / 194.71
Min FPS: 78.62 / 81.66 / 83.42
Frame times (3-run average): Avg. 7.81 - Lower 1% 11.1 - Lower 0.1% 13.7
Shadow of Mordor is tied in performance with the previous driver. FPS are on the same ranges, and Frame time numbers recorded are maybe a hair better, but well within error margin again.
Driver testing conclusions
System stability is good with the new driver. For now, The Division (except for a reproducible crash to desktop during benchmark which I already know how to bypass), Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor and WoW (short testing game sessions) all ran fine without stability issues.
I’ve noticed though that something has changed with the brightness and contrast balance of the image after installing the new driver (not only in games, but everywhere, including the Windows desktop). On my particular scenario, this makes the image look better than the previous driver. Better contrast balance, and a more brighter image, (which allowed me to lower a bit the brightness setting of my monitor that was set to maximum before).
But be aware that some people in this same reddit thread also noted differences, but in the opposite way, with worse image quality and white balances (while other people do not report any change at all). Seems the only way to find out will be testing the drivers on your machine, and roll the dice ;)
Performance wise, it’s safe to say that this driver does not change things for Pascal owners. Most games do not see any noticeable performance change, and while average FPS on FarCry 5 may be a hair lower, lower 1% and 0.1% frame times are a tiny bit better, thus indicating better image stability and less stuttering.
Tested also Monster Hunter: World. While I don’t take numbers for this one as it lacks an ingame benchmark tool, I haven’t noticed any change on the game after the new driver.
My recommendation
Summing it up. Once again seems safe to upgrade, no significant changes in game performance on my setup, but you will have to be careful about the contrast/brightness changes that seems to happen on certain configurations with this driver. It may be good for you, bad for you, or you may not notice anything at all.
417.01 is the Game Ready driver for Darksiders 3
416.xx version
I have created no summary graphs or tables for the 416.xx version due to some changes to the benchmark PC in 416.16 and 416.34.
Driver 416.94
Tom Clancy’s: The Division
Three consecutive runs with 416.94:
Avg. FPS: 86.2 / 86.3 / 86.3
Typical FPS: 86.7 / 86.7 / 86.8
Avg. CPU: 69% / 66% / 68%
Avg. GPU: 96% / 96% / 96%
Performance identical to previous drivers, as is the subjective smoothness perception during benchmark. No changes at all.
Ghost Recon: Wildlands
Three consecutive runs with 416.94:
Avg FPS: 80.18 / 79.22 / 79.38
Min FPS: 68.79 / 70.09 / 69.67
Max FPS: 91.27 / 89.46 / 89.98
Avg CPU: 51.3% / 51.2% / 51.4%
Avg GPU: 96.6% / 96.7% / 96.6%
Frame times (3-run average): Avg. 12.7 - Lower 1% 15.9 - Lower 0.1% 20.5
Wildlands is also almost on par with previous 416.81. Maybe 1 extra fps up on average, but may very well be test variance.
Subjective game smoothness is also identical, no noticeable stutters at all.
FarCry 5
Three consecutive runs with 416.94:
Min FPS: 70 / 73 / 71
Avg FPS: 88 / 90 / 89
Max FPS: 109 / 111 / 109
Frame times (3-run average): Avg. 11.1 - Lower 1% 17.0 - Lower 0.1% 22.9
Another carbon copy of the previous results. Same FPS numbers, no stutters, same smoothness perception on the game.
Batman: Arkham Knight
Three consecutive runs with 416.94:
Min FPS: 41 / 40 / 42
Max FPS: 122 / 123 / 122
Avg FPS: 84 / 84 / 83
Frame times (3-run average): Avg. 12.1 - Lower 1% 20.8 - Lower 0.1% 26.5
Arkham Knight doesn’t show any changes at all since 416.81.
Middle Earth: Shadow of Mordor
Three consecutive runs with 416.94:
Avg FPS: 131.89 / 130.88 / 131.38
Max FPS: 195.69 / 191.17 / 192.20
Min FPS: 86.28 / 84.58 / 85.32
Frame times (3-run average): Avg. 7.83 - Lower 1% 11.3 - Lower 0.1% 14.0
Shadow of Mordor is maybe 1 FPS down on average, which is very well under the error margin of the test. To all intents and purposes performance seems the same as previous driver.
Driver testing conclusions
System stability is good with the new driver. The Division (except for a reproducible crash to desktop during benchmark which I already know how to bypass), Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World, Middle Earth: Shadow of Mordor and WoW (short testing game sessions) all ran fine without stability issues.
Performance wise, this driver is identical to the previous one under Pascal cards and for old games, as expected given the small version jump and the tiny timeframe between driver releases. Seems Nvidia is just releasing this for improving performance on Battlefield V (RTX is now working and Nvidia need to squeeze as much performance as possible for showcasing the new cards) and for the new GameReady releases.
Tested also Monster Hunter: World. No changes neither on this one on my PC.
My recommendation
Summing it up. Seems safe to upgrade, but you won’t notice any change unless playing the newest Game Ready releases (BFV seems to be working much better now under DX12 though).
Again, if you are still running the old 3xx drivers, and you don’t have any issue on your current games, this new driver still does not offer anything new for you.
416.94 is the Game Ready driver for Hitman 2
Driver 416.81
Tom Clancy’s: The Division
Three consecutive runs with 416.81:
Avg. FPS: 85.8 / 86.2 / 86.1
Typical FPS: 86.2 / 86.7 / 86.1
Avg. CPU: 68% / 64% / 64%
Avg. GPU: 96% / 96% / 95%
Performance dropped by a couple of Frames per Second. Nevertheless, I have to say that this drivers ideed fixed some strange geometry corruption/LOD detail levels I had on the ingame benchmark since The Division DX12 patch was released (even on my old 970 GTX).
This corruption was clearly visible on the background bridge which is rendered during the last segment of the benchmark, so I assume the lost performance is indeed because now that geometry is being correctly rendered (and the bridge looks gorgeous) while it was not being rendered properly before the driver update (it was a polygon mess).
Subjective smoothness and frame pacing perception during benchmark and gameplay stays in the same level than the previous driver.
Ghost Recon: Wildlands
Three consecutive runs with 416.81:
Avg FPS: 79.98 / 78.94 / 78.60
Min FPS: 68.05 / 70.23 / 69.17
Max FPS: 90.55 / 89.11 / 88.21
Avg CPU: 51.4% / 51.9% / 55.4%
Avg GPU: 96.5% / 96.4% / 96.4%
Wildlands, on the other hand, seems to have slightly better numbers on average. Gains are so small that they are within the error margin of the test, but the fact that every single number is marginally better is a good symptom.
Subjective game smoothness is great on both drivers. No noticeable changes here.
FarCry 5
Three consecutive runs with 416.81:
Min FPS: 69 / 73 / 68
Avg FPS: 87 / 89 / 88
Max FPS: 110 / 111 / 108
Same result on FarCry 5 with this new drivers than with the previous ones. Numbers might be marginally better again, but so small that may be just test variance.
Smoothness perception and frame pacing identical to 416.34.
Batman: Arkham Knight
Three consecutive runs with 416.81:
Min FPS: 23 / 40 / 41
Max FPS: 123 / 127 / 125
Avg FPS: 83 / 85 / 84
Except for a weird 23 minimum FPS spike on the first test of the new driver (which I consider an outlier and not significant, as the average was exactly as expected), there are no changes at all with the new drivers for Arkham Knight.
Middle Earth: Shadow of Mordor
Three consecutive runs with 416.81:
Avg FPS: 132.23 / 131.91 / 132.09
Max FPS: 190.37 / 191.78 / 189.94
Min FPS: 86.80 / 87.43 / 87.10
Shadow of Mordor show more or less the same behavior than Ghost Recon Wildlands. Small gains which are not very significant, but consistent between runs. Averages up by 1.5-2 FPS, same maximum, and slightly better minimum FPS too.
Driver testing conclusions
System stability has been good all around. I can consistently reproduce a crash-to-desktop when running the DX12 test of The Division (like happened to me during the previous driver tests), but now I’ve learned how to work around it.
The rest of my usual test games ran fine: Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor (short testing game sessions).
Performance wise, the driver seems stable too. The Division performance may have worsened a bit under DX12, probably due to the fixes included in the patch to avoid certain graphical glitches, but in turn DX11 games are on the same level of performance as under 416.34, or even a hair better all around.
My recommendation
As a conclusion, given the huge amount of bug fixes included, (and the launch of Battlefield V) if you are already running a driver of the 411/416 branch I think it could be worth upgrading to this latest driver.
If you are still running the old 3xx drivers, (and you don’t have any issue on your current games), this updated drivers doesn’t seem to offer anything new for you, so you may very well keep yours.
Also, a final note for those playing hard Monster Hunter World, (like myself ;) ): I’ve not played too much with MH:W and the new drivers, and I don’t have numbers to prove it, but I’d swear the game is now running smoother than before. Just hunted a Teostra on Elder’s (one of the worst scenarios performance wise), and the hunt went butter smooth.
416.81 is the Game Ready driver for Battlefield V
Driver 416.34
Note: The Benchmark PC used for driver 416.34 and older versions had 8GB of ram.
416.16 driver performance losses with Win10 v1809 patch installed
Wildlands three consecutive runs with 411.70 (W10 v1803 April Update, no Spectre patch installed yet):
Avg FPS: 73.22 / 73.13 / 73.96
Min FPS: 64.87 / 66.00 / 66.34
Max FPS: 85.00 / 83.09 / 83.25
Avg CPU: 73.4% / 74.1% / 73.7%
Avg GPU: 90.1% / 88.3% / 89.4%
Wildlands three consecutive runs with 416.16 (W10 v1809 October Update, which installed Spectre microcode patch):
Avg FPS: 70.49 / 71.05 / 69.69
Min FPS: 54.78 / 58.88 / 59.41
Max FPS: 80.52 / 79.76 / 79.00
Avg CPU: 77.7% / 74.3% / 78.0%
Avg GPU: 84.8% / 86.2% / 82.3%
Wildlands three consecutive runs with 416.16 (W10 v1809 October Update, but Spectre patch disabled):
Avg FPS: 72.82 / 73.01 / 72.97
Min FPS: 63.81 / 62.10 / 64.45
Max FPS: 84.90 / 85.09 / 83.51
Avg CPU: 69.8% / 71.1% / 70.9%
Avg GPU: 92.3% / 91.2% / 92.1%
With this numbers, I can confirm that most (if not all) the performance I lost on the previous driver test was not because the new Nvidia drivers, and not because the Windows 10 v1809 October update itself, but due to the Spectre microcode patch (which was installed in my machine along the v1809 October update). Some people already got it installed in the past as a standalone KB4100347 update, so they won’t probably notice any difference after Win10 v1809 October update.
Note that the Meltdown and Spectre patches do not have any direct effect on your GPU nor your graphics horsepower, only over your Intel main processor. You won’t see any difference in performance unless your gaming is somewhat CPU bottlenecked (as was my case with the mid range i5-4590 4th gen processor with a pretty powerful 1070Ti GPU under DX11 games).
If your system is mostly GPU constrained (low-mid end GPU or a high end main processor), the Spectre microcode patch shouldn’t affect your gaming performance.
Tom Clancy’s: The Division
Three consecutive runs with 416.34:
Avg. FPS: 87.7 / 87.6 / 87.6
Typical FPS: 88.3 / 88.2 / 88.1
Avg. CPU: 69% / 66% / 69%
Avg. GPU: 96% / 96% / 95%
Once again performance doesn’t change at all in the Division. The same FPS numbers, and the same perceived smoothness (stuttering almost non-existant).
Ghost Recon: Wildlands
Three consecutive runs with 416.34:
Avg FPS: 78.41 / 78.34 / 78.48
Min FPS: 59.46 / 68.52 / 68.18
Max FPS: 89.46 / 88.82 / 88.65
Avg CPU: 48.5% / 48.0% / 48.4%
Avg GPU: 95.9% / 96.2% / 92.1%
Wildlands stayed the same on the new 416.34 drivers too. Just only slightly lower average CPU usage (which is good, given the steady performance).
Smoothness is great on both drivers. Once I updated my CPU to i7, those lag spikes I got on my previous test during fast driving scenes went away (coherent with the increased CPU bottleneck I had in my old i5 due to the Spectre patch).
Notice how the performance here went up by a whooping 10% just by upgrading the i5 CPU to the i7. Again we can see how DX11 games are much more reliant on CPU horsepower than DX12 games.
FarCry 5
Three consecutive runs with 416.34:
Min FPS: 69 / 72 / 70
Avg FPS: 86 / 88 / 88
Max FPS: 108 / 109 / 109
FarCry 5 performance seems to have improved a small bit with this driver. Minimum, maximum and averages are consistently up by a couple of frames per second.
Smoothness perception and frame pacing is maybe a hair better too with 416.34.
(As a side note, this is another DX11 game in which I’ve got more than 10% extra performance just by upgrading the CPU).
Batman: Arkham Knight
Three consecutive runs with 416.34:
Min FPS: 42 / 43 / 43
Max FPS: 123 / 124 / 123
Avg FPS: 82 / 83 / 85
No changes at all with the new drivers and Arkham Knight.
Middle Earth: Shadow of Mordor
Three consecutive runs with 416.34:
Avg FPS: 130.95 / 130.93 / 130.83
Max FPS: 191.45 / 191.34 / 192.30
Min FPS: 73.50 / 85.06 / 83.17
And once again, Shadow of Mordor does not have any noticeable changes on the new driver.
Driver testing conclusions
System stability has been good all around. I’ve got another crash-to-desktop while testing The Division, but it seems that this is pretty common under Pascal architecture cards and DX12 since it got released a couple of years ago and under all driver branches.
The rest of my usual test games ran fine: Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor (short testing game sessions).
My recommendation
Performance wise, this driver is stable. No significant gains or losses compared to 416.16, except maybe the couple of FPS up on FC5, which is always welcome.
Some people are claiming some flickering issues on different scenarios, (and even the patch notes acknowledge them on FC5), but I’ve not yet found this bug myself.
Also, for those playing Monster Hunter:World like myself, I’ve not noticed any difference before and after the driver update. ;)
416.34 is the Game Ready driver for Call of Duty: Black Ops 4.
Driver 416.16
Note: The Benchmark PC used for driver 416.16 and older versions had a different CPU, i5-4590.
Here you have my early performance benchmark, this time for the new 416.16 WHQL plus Windows 10 October’s Update (v1809), compared to the previous 411.70 driver release. Big changes on the OS, plus a significant jump in Nvidia version number.
Tom Clancy’s: The Division
Three consecutive runs with 416.16:
Avg. FPS: 87.4 / 88.0 / 87.3
Typical FPS: 87.8 / 88.5 / 87.7
Avg. CPU: 106% / 106% / 107%
Avg. GPU: 95% / 95% / 95%
Performance haven’t changed here at all. The same FPS numbers, and the same perceived smoothness and stuttering (almost non-existant I must say).
Notice the Average CPU count on the new tests... 106%-107% I assume this should be a bug somewhere between The Division code and the new Windows 10 update.
Ghost Recon: Wildlands
Three consecutive runs with 416.16:
Avg FPS: 70.49 / 71.05 / 69.69
Min FPS: 54.78 / 58.88 / 59.41
Max FPS: 80.52 / 79.76 / 79.00
Avg CPU: 77.7% / 74.3% / 78.0%
Avg GPU: 84.8% / 86.2% / 82.3%
Wildlands got a noticeable drop in performance on my rig of about 3-4 FPS (about a 5% loss), and the Minimum FPS went down by even more. What is worse, I’ve found some stutterings on fast action scenes that weren’t there on the 411.xx drivers.
Also, I’ve noticed that average CPU usage went up a bit, while GPU usage went down. This might indicate a CPU bottleneck, so maybe the drivers are not slower by themselves, but they might be using more CPU power on DX11 games (at least under the new W10 1809 update).
Those with new and fast processors might not notice any changes, but those with limited horsepower on the main CPU might see their performance reduced.
FarCry 5
Three consecutive runs with 416.16 (W10 v1809):
Min FPS: 56 / 58 / 58
Avg FPS: 77 / 74 / 75
Max FPS: 102 / 96 / 99
Like Wildlands, FarCry 5 took a noticeable dive on performance. Averages are 6-7% down, and minimum FPS are lower too, but by a smaller amount. Smoothness perception and frame pacing is worse now than the previous 411 driver with W10 April’s update.
Driver testing conclusions
Well... This one has been a small disappointment for me.
I’ve been reading quite good things about the new Windows 10 October’s Update. Also, the previous 411 driver branch (and the very last release of the 399 branch) were very good for Pascal owners, with improved FPS performance, lowered stuttering, and provided better smoothness all around, so I believed this one could follow the same trend.
Unfortunately, while The Division in DX12 mode is performing about the same as the previous drivers, both DX11 games are now running about 5-6% slower, quite a serious performance hit. What is worse, I’ve noticed lags and stutters in places that were smooth before. Frame pacing seems worse overall too.
I don’t know if the drivers are to blame, or if it’s the Windows v1809 update (or maybe the combination of both updates mixed together), but whatever the cause may be, now DX11 games are performing noticeably worse on my machine.
Update: Thanks to reddit user LazyGit, I’ve found that the Spectre mitigation patch for Intel processors, which is known to increase CPU usage and lower performance a bit, got installed on my machine along the W10 v1809 October update. This may very well explain the increased CPU usage and thus the lower FPS performance.
My system is pretty much CPU constrained, other systems with more breathing room on the CPU side might not see this performance drop.
System stability has been problematic too. I’ve got a crash-to-desktop while testing The Division, and a complete system crash and reboot while running XCOM 2. The rest of my quick “cheat sheet” suite worked fine, at least for now, (Wildlands, FarCry4, FarCry5, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor during short testing game sessions).
But it has been QUITE a while (I mean, probably a year or more) since I got a game crash on my system for the last time.
My recommendation
In the end, for now I cannot recommend the Windows 10 October’s Update or the new 416 driver branch. Without knowing the exact cause of the lost performance, I’m not able to give a firm opinion on either update.
To make matters worse, with the Windows 10 Spectre microcode mitigation patch for Intel processors installed on my PC at the same time, another extra variable has been added to the mix.
This Windows Spectre patch may very well explain the difference in performance, thus “exonerating” both W10 and the Nvidia drivers from the lost performance. But there is no way I can remove just that part of the October’s Update and check the rest.
I will have to wait to the next driver release to see the performance trend by then.
The total number of upcoming DLSS games is 25 - Nvidia
411.xx version
Driver 411.70
Tom Clancy’s: The Division
Three consecutive runs with 411.70:
Avg. FPS: 87.3 / 87.8 / 87.9
Typical FPS: 87.7 / 88.2 / 88.3
Avg. CPU: 84% / 88% / 89%
Avg. GPU: 95% / 95% / 95%
A small performance regression here. It may just be test variance, as then change is very small and well within error margin. About 1 FPS less on average on the three test runs. Stuttering and subjective game smoothness is identical to the previous driver release (411.63 was very good for my setup in this regard, so it’s fine).
Ghost Recon: Wildlands
Three consecutive runs with 411.70:
Avg FPS: 73.22 / 73.13 / 73.96
Min FPS: 64.87 / 66.00 / 66.34
Max FPS: 85.00 / 83.09 / 83.25
Avg CPU: 73.4% / 74.1% / 73.7%
Avg GPU: 90.1% / 88.3% / 89.4%
This one is a carbon copy compared to the previous drivers. Numbers are almost the same in all metrics. Previous driver got a small but noticeable bump in performance, and it is maintained in this driver. Game smoothness and micro-stuttering in line with the FPS numbers, in the same trend as the excellent previous release. No changes here, which is good news :)
FarCry 5
Three consecutive runs with 411.70:
Min FPS: 60 / 62 / 61
Avg FPS: 80 / 80 / 79
Max FPS: 103 / 105 / 105
Again, FarCry V have almost the same numbers as the previous driver (which got a couple of bumps in the last releases so it’s ok). Maybe even 1-2 FPS above compared to previous 411.63, but again change is so small that it may be just test variance. Smoothness and stuttering is very good, also in line with the last driver.
Driver testing conclusions
As this is a very small version jump, I didn’t expect big changes. And indeed, things went as expected. Maybe a very small drop on The Division, exact same numbers on Wildlands, and a very small bump on FC5. Overall, it’s safe to say that performance didn’t change in any meaningful way this driver (for Pascal owners at least; I guess Nvidia should still be tinkering with Turing changes and the 20xx RTX lineup, so they may have noticed more changes than us).
System stability has been good so far. For now, I’ve got no game crashes nor driver restarts. My standard picks for stability testing went smooth. The Division, Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor all tested and working fine (short testing game sessions).
For those playing Monster Hunter World, I haven’t noticed any change at all since 411.63. Performance got improved a bit in the previous driver, and is still the same on this one.
Some players on 10xx cards reported that they have got some flickering with the game, but I’ve not been affected by this. Also, seems that some people is again getting improved performance by manually deleting the Monster Hunter World Nvidia default profile using the Nvidia Profile Inspector tool, but for me the game is now stable enough, so I haven’t messed with this since 411.63.
My recommendation
So, in the end, this is a pretty dull driver release for 10xx Pascal owners. No noticeable performance changes, and only a handful of fixes, mostly for Turing owners. The only interesting things - in my opinion - are the new Game Ready games (but I’m not going to play any of them anyway), and this small but annoying bug fixed -> “Custom color settings reset on reboot in Windows 10 FCU”.
411.63 added support for the new GeForce RTX GPUs.
Driver 411.63
Tom Clancy’s: The Division
Three consecutive runs with 411.63:
Avg. FPS: 88.3 / 89.2 / 89.0
Typical FPS: 88.7 / 89.6 / 89.4
Avg. CPU: 89% / 88% / 92%
Avg. GPU: 95% / 95% / 96%
A small but consistent increase in framerate. Nothing stellar, but significant above error margin. Also, subjective smoothness seems better overall during the benchmark, and micro-stuttering seems less frequent. During gameplay I’m playing capped to vSynced 60 fps, and smoothness is perfect in any case without any noticeable stutter at all.
Ghost Recon: Wildlands
Three consecutive runs with 411.63:
Avg FPS: 73.72 / 73.64 / 73.29
Min FPS: 64.23 / 64.74 / 61.64
Max FPS: 83.09 / 84.24 / 82.84
Avg CPU: 73.5% / 73.9% / 73.8%
Avg GPU: 89.1% / 89.0% / 88.5%
Once again I’m getting another noticeable improvement over the previous driver (and remember that 399.24 was in turn a huge jump over the precedent 399.07). Average framerate increased by about 3-4 extra FPS (5%), with minimum and maximum numbers increasing too. Game smoothness and microstuttering was greatly improved in the previous 399.24, and once again this benchmark seems to run even smoother.
FarCry 5
Three consecutive runs with 411.63:
Min FPS: 59 / 59 / 60
Avg FPS: 78 / 77 / 78
Max FPS: 102 / 99 / 101
Far Cry V have almost the same numbers than the previous driver. Same Max and Min, and maybe 1-2 extra Average FPS, but should not be considered significant as it’s well within the test error margin. Smoothness and stuttering also in line with 399.24.
Driver testing conclusions
Before stating my conclussions, I have to say that I had to test all readings twice. My first tries not only gave me lower numbers on DX11 games (in some cases by quite a big amount), but also got a few nasty “back to desktop” crashes in games, and even one bluescreen Windows 10 crash.
Worried as drivers were the only change I made, I remembered that in the rush I installed the new ones over the previous ones using Nvidia Experience without uninstalling 399 first, so performed a Safe Mode clean DDU uninstall, and installed the new drivers from scratch.
Indeed, the clean reinstall seems to have cleaned something that was messed up during the regular update process, because after the clean reinstall not only the crashes are completely gone, but also both DX11 games have much better readings.
As this is a big branch jump (which should probably contain quite a number of changes under the hood), my first advice is to perform a clean DDU reinstall if you run into any issue or noticeable FPS loss.
Now, on the test itself. After the clean reinstall system stability has been good so far, no more game crashes nor driver restarts. The Division, Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor all tested and working fine (short testing game sessions).
Performance wise, this driver is again pretty interesting. On my Pascal 1070Ti both The Division (DX12) and Wildlands (DX11) got improvements in framerate and frame pace smoothness, and FarCry 5 is almost on par or even a hair better than before. Seems that once again CPU usage has improved, so this can explain the gains on my rig with a midrange 4th gen. i5.
People less constrained by CPU power with better processors maybe won’t see much of a change, but if you have a mid range/old CPU, there is a chance that this drivers can help you with it.
My recommendation
Coupled with the tons of new features, updated game profiles, and quite a few old time bugs fixed, this driver package seems a must for Pascal owners. No matter which driver you came from, this time I’d give the new ones a try regardless. You can always DDU uninstall and go back to the previous one if you don’t like what you see.
Of course, given that this release is a big branch jump with many changes, it’s not unexpected that some issues might arise in the coming days-weeks. But so far, I’m pretty happy with the drivers and they seem pretty solid (but only after the Clean Reinstall with DDU cleanup, as the first regular “overwrite” install over 399.24 gave me quite some headaches)
For those playing Monster Hunter World. Performance seems much much better with this driver. Elder’s Recess and Rotten Vale are working smoother than before, and even as I’m not yet able to fight Teostra or Kushala Daora, the fights with Kirin and Legiana also improved noticeably (probably due to improved lighning/ice effects respectively)
The Nvidia Profile Inspector trick seems that is no longer needed.
399.24 is the Game Ready driver for Shadow of the Tomb Raider.
399.xx version
Driver 399.24
Tom Clancy’s: The Division
Three consecutive runs with 399.24:
Avg. FPS: 87.4 / 87.2 / 87.9
Typical FPS: 87.8 / 87.5 / 88.3
Avg. CPU: 84% / 86% / 86%
Avg. GPU: 95% / 95% / 95%
Game numbers are almost identical to the previous drivers. Noticed a few % decrease in CPU usage, which is good (more on this later). No changes on framerate. Subjective smoothness and micro-stuttering seems also in the line of previous release. During gameplay I’m always capped to 60 fps vSynced with both drivers, so smoothness is almost perfect in any case without any noticeable stutter.
Ghost Recon: Wildlands
Three runs with 399.24:
Avg FPS: 70.90 / 68.96 / 70.59
Min FPS: 60.28 / 61.08 / 57.43
Max FPS: 80.76 / 75.62 / 78.53
Avg CPU: 76.2% / 76.5% / 76.8%
Avg GPU: 85.5% / 83.0% / 86.7%
This time I’m getting quite a noticeable improvement over the previous driver. Average FPS increased by a whooping 8.8%, with minimum and maximum numbers increasing too by a varying degree. Not only that, but stutters are much less noticeable, and benchmark ran much smoother.
On 399.07 I had <80% GPU usage so my GPU was running slower than it could, and given the almost constant +-80% CPU usage I may had some CPU bottleneck. With newest 399.24 the Average CPU usage went down, while Average GPU usage went up. This could very well explain the improvement in framerates and game smoothness (details on this below).
FarCry 5
Three runs with 399.24:
Min FPS: 60 / 60 / 60
Avg FPS: 77 / 76 / 77
Max FPS: 102 / 98 / 104
And again the runs with the new drivers have an increase in Far Cry 5 performance of almost 10%. Numbers are consistent in Max, Min and Averages. I ran too a few more FC5 benchmark runs just to be extra-sure, and numbers are almost the same as the ones I captured for this post. The increase does exist.
Driver testing conclusions
First of all, system stability has been good so far. The new drivers seem solid. No crashes nor driver restarts. The Division, Wildlands, FarCry4, FarCry5, XCOM2, EVE: Online, Dauntless, Terraria, World of Tanks Blitz, Batman Arkham Knight, BattleTech, the Mass Effect trilogy, Monster Hunter: World and Middle Earth: Shadow of Mordor all started and played without issues (short testing game sessions).
Now on the test result itself:
First impression I had on this driver was quite amazing. A few bugfixes, new GameReady games, and performace went through the roof on DX11 games by up to 8-10%. A must have!
But after thinking quite a bit about it, at this point the difference is so huge, and my rig have been just upgraded a couple of days ago, so I’m not sure this change is related to the updated Nvidia driver TBH.
I’m starting to believe that my new GTX 1070Ti setup is CPU constrained on 1080p DX11 games even with high demanding graphics settings enabled, and something was running in the background of my Windows 10 installation without my knowledge while I ran the 399.07 benchmarks, yielding lower numbers than expected. DX12 games are known for using less CPU, so The Division in DX12 mode ran without noticeable changes (besides a slightly lower CPU usage, which is consistent with the other games result).
Or, it could be that the new Nvidia drivers do indeed improve by lowering CPU usage, so when the CPU is holding back the performance we get better numbers overall. Until I find out what happened with my 399.07 numbers, I cannot give a straight recommendation performance-wise.
My recommendation
Stability-wise the drivers are solid, installed fine, and they run fine on my PC without crashes or restarts.
Performance seems at least as good as previous driver (if no better as discussed), so for the time being my recommendation stay the same as with the previous driver releases of the 397/398/399 branch: If you are still on Nvidia 391.xx or previous driver branches and your games are working fine, you can keep your drivers, but on any other scenario I would download and install this driver release.
If you have a powerful GPU coupled with a mid-range or old CPU though, I encourage you to upgrade ASAP anyway. This driver may have (or not) a noticeable effect by lowering the CPU dependence. If you are in this scenario, please share here your experiences too.
Header and all the images in the article: Nvidia
