#6017016 Posted on: 05/14/2022 05:22 AM

---

I'm all happy with these drivers

#6017033 Posted on: 05/14/2022 09:34 AM

---

Using 472.84 because it's the smoothest in Unreal Engine games on my GTX 1070.

512.59 was the same stuttery mess like all previous drivers after 472.84. So until i get a RTX 4000 i will stick to this old driver.

#6017082 Posted on: 05/14/2022 12:45 PM

---

Objective benchmark analysis of drivers from 442.59 to 512.77 using custom spreadsheet I originally created to objectively analyse raw benchmark data (see NOTE in footer).

Raw benchmark data on the left is from historical and recent posts by LokkenJP over on Reddit (1070ti GPU, i7 4790K CPU, Windows 10).

Summary columns to the right objectively highlight the best driver for a given category (FPS, frametimes AVG 1% 0.1%) in the games tested. Each driver is given a point / score based on how many times it is selected as the 'best' for a given game and category (see NOTE in footer).



NOTE: Spreadsheet 'summary' columns automatically select the 'best' driver for each game in each of the 4 categories ( FPS, frametimes AVG 1% 0.1% ). If a category result 'matches' across multiple drivers the higher driver number in the result is selected e.g. if 442.49 and 512.59 have matching FPS results then 512.59 is selected in the 'summary' list. FPS figure is the MEDIAN from the 3 raw benchmark data points.

#6017113 Posted on: 05/14/2022 04:41 PM

---

Objective benchmark analysis of drivers from 442.59 to 512.77 using custom spreadsheet I originally created to objectively analyse raw benchmark data (see NOTE in footer).

Raw benchmark data on the left is from historical and recent posts by LokkenJP over on Reddit (1070ti GPU, i7 4790K CPU, Windows 10).

Summary columns to the right objectively highlight the best driver for a given category (FPS, frametimes AVG 1% 0.1%) in the games tested. Each driver is given a point / score based on how many times it is selected as the 'best' for a given game and category (see NOTE in footer).



NOTE: Spreadsheet 'summary' columns automatically select the 'best' driver for each game in each of the 4 categories ( FPS, frametimes AVG 1% 0.1% ). If a category result 'matches' across multiple drivers the higher driver number in the result is selected e.g. if 442.49 and 512.59 have matching FPS results then 512.59 is selected in the 'summary' list. FPS figure is the MEDIAN from the 3 raw benchmark data points.
Good work u did there. But not to be annoying every system acts different on every driver its never the same like 512 may work for u better while other prefer 496 or so.

#6017116 Posted on: 05/14/2022 04:53 PM

---

Objectively speaking, I have reported the display corruption that takes place on DCH, Appx / MS Store issues since the incept of 496. Those are well documented throughout the drivers thread. I also have 4 tickets in.

this is a specific use case: using 3090 and 3090 ti cards, in 10 bit, 4k, solid color background, HDMI 2.1 or DP 1.4 DSC. running at 120hz in DP 1.4 and HDMI 2.1 or 144 in HDMI 2,1- windows shows black small polygons on the 2d desktop and around open windows. If you let windows do the internal screenshot - the polygons are in the screenshot so they are rendered in windows and are not hardware (card / monitor).
reproducible on multiple PCs and monitors with the same common scenario,

I'm not discussing performance, this is strictly a visual corruption anomaly that goes away when appx is set to disable and MS Store is removed from the build in windows server 2022 and windows 11. Here and in 4 other forums, over time, many users have reported the issue and provided their own screenshot of exactly the same manifestation.

So actually yes, MS Store and DSC do cause issues, from a specific use scenario.
Issue 100% resolved by either reverting to a standard driver or disabling Appx/MS store


This is a screen shot of the 1-3 mm polygons. Its a screen shot, so windows is rendering the polygons, (not a hardware issue).

11228