Page 1 of 2

Performance difference between 425.31 and 452.06

Posted: Mon Sep 28, 2020 10:27 am
by ZeGURU69
Hello,

I tried to update my 3Dvision driver on windows 7 from 425.31 to 452.06 via 3Dfix Manager and I noticed a noticeable drop in performance. On games like Divinity 2, Borderland 3 or Vampyr I have at least 30% less performance.

I also have the 452.06 driver on Windows 10 (1809) and it runs like Windows 7. I have an RTX 2080 Ti. Someone has already noticed that the performance drops on the drivers after the 425.31, is there a manipulation to be done to have the same performance ?

Thank you for your answers !

Re: Performance difference between 425.31 and 452.06

Posted: Mon Sep 28, 2020 10:34 am
by ZeGURU69
I add that I am of course on standard drivers on Windows 7 and 10 and not driver type DCH

Re: Performance difference between 425.31 and 452.06

Posted: Thu Oct 01, 2020 4:28 pm
by ZeGURU69
Nobody has the problem or is it normal ? I just reinstalled the 425.31 driver on Windows 10 and some games run better. Then I reinstall the 452.06 drivers via 3Dfix manager and again the games have a slower framerate.

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 4:17 am
by russellk
Are you saying the performance drop is only on Windows 7 or on Windows 10 (1809) as well? Also, are you talking about 2d, or 3d or both?
Lastly, how are you measuring performance?

I can't help directly as I'm leaving my drivers alone for now, but something is wrong as it's the first I've heard of this. Are you sure it's not down to Vsync settings or power settings?

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 5:26 am
by valery121
https://aduk.de/ has the problem or is it normal ? I just reinstalled the 425.31 driver on Windows 10 and some games run better. Then I reinstall the 452.06 drivers via 3Dfix manager and again the games have a slower framerate.

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 8:55 am
by ZeGURU69
The performance drop on Windows 7 and Windows 10 with 452.06 and only about 3D. In 2D I haven't seen any difference

For performance I run Divinity 2 which is cap at 60 fps on 425.31. On the 452.06 with my savegame I am at 45 fps, if I go back to 425.31 I go back to 60 fps. Borderland 3 does a bit of the same thing (except that even on 425.31, it drops below 60 fps).

Basically all the games that are a bit greedy have a lower framerate on 452.06 compared to 425.31 in 3D whether on windows 7 or windows 10. My processor is an I7 8700K@4.9GHZ

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 2:46 pm
by ZeGURU69
Graphics settings are exactly the same Vsync enabled and power optimized for both driver.

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 3:21 pm
by bo3bber
Are you using HelixVision and an HMD? That 45 fps cap is suspiciously identical to the frame rate limit we apply for the VR Display Profile.

You can also check the nvidia control panel frame rate limiter. It only exists in the later drivers, not in 425.31.

Re: Performance difference between 425.31 and 452.06

Posted: Fri Oct 02, 2020 3:57 pm
by ZeGURU69
No, I am not using HelixVision and my framerate is not cap at 45 fps. Divinity 2, I saved in a place where it is more graphically greedy but in other places I am at 60 fps on 452.06. Borderland 3 I can reach 60 fps on the 452.06 but much less often than on the 425.31. The frame rate limiter is indeed disabled

I really don't know why there is a difference in framerate and I'm surprised no one has had this problem.

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 03, 2020 10:53 am
by ZeGURU69
I tried the losti method to install the 452.06 by going through DDU and installing the 425.31 first but I still have the same results ...

Re: Performance difference between 425.31 and 452.06

Posted: Mon Oct 05, 2020 10:26 pm
by Lizzard
For what my 50 cents are worth.
On 1809 I had CPU problems. When I booted my PC fresh, instead of waking it up from a sleep, CPUID HWMonitor showed the boost executed properly. 4990Mhz. When I woke it up from a sleep the frequency would have been back to stock 3600Mhz. So I had to use ThrottleStop to reset the boost every time I woke it. 1809 in my experience is not good for your CPU. (or at least you're not getting the most out of it) Especially on later Geforce drivers. I do believe later versions are better optimized for your CPU in conjunction with latest GPU drivers. (depending how recent your CPU is) I personally have not jumped back to 1809 since 1903 got forced down on me after a year's use. The difference was quite substantial and have to say the transition was pretty seamless. System just feels a lot more stable, responsive and harmonious.

I have to mention that for the most part I used an i7 3820 during my year on 1809. Pretty old CPU, which was quite expensive at launch. And the performance was not too bad. Games are still about 67% to 90% GPU bound. If I had to generalise. Which is why Rise of the Tomb Raider never lagged on this CPU with a 1080ti. It only really started to lag with Shadow of the Tomb Raider. This might be an exception since SOTTR is a bit more CPU bound according to benchmarks. If you are one of those who don't mind to reinstall windows often then try something later or newer. After being on W8.1 for about 5 years i really hated the transition to W10. W10 continually tries to strong arm you into updates and stuff. As if the system doesn't serve you anymore. As if you are serving the system. And I really miss W8.1. It was such a humble OS. W10 is such an arrogant my way or the high way experience on top of its BSOD tantrums. Really pisses me off. Windows Mixed Reality is the only reason why I'm using it. Once the VR bug bites, you might not want to look back. The guys who never experienced 3d vision on a projector might not be as forgiving as those who did. But having a massive projector screen clamped onto your face, that is bigger than anything you could practically make for you personal use, where you really could juice the perfection out of the pixels, has never been more immersive. One of the fundamental principles of being immersed is to have a completely dark environment where it is only you and the pixels and the sound. It embraces you. And it doesn't let go easily.

As of latest reports, 3d Vision and 3dmigoto is not that finicky about the Windows version anymore. (with the 30 series and its drivers out of account). Team Helix really overcame many obstacles to date. So unless someone else could give a good reason? why do you want to stick with 1809?

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 10, 2020 5:22 am
by ZeGURU69
Hello,

Thank you for that answer. I stay on version 1809 because I'm like you, I don't like updates being forced and I don't need a more recent version of windows. I've tested VR before but I don't like having a headset on my head while playing.

I have some doubts that version 1809 restricts the performance of the latest Nvidia drivers in 3D. If it did, there should be a performance drop in 2D as well. Also, I'm not the only one using an older version of windows and no one has mentioned a drop in performance in 3D.

Finally, I would have to do a fresh installation of windows 2004 to see if the performance differences are still there. I am amazed that no one has seen this drop in performance with drivers after 425.31 whether under Windows 1809 or Windows 7

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 10, 2020 4:22 pm
by Necropants
Something is funny there. I run 1809 on the lastest driver where it still works and if anything I have better performance.
Sure you used DDU propery in safe mode?

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 11, 2020 12:14 am
by Lizzard
I didn't notice worse performance either. But definitely some stability issues that are unique to my system. There are some extra bells and whistles in later drivers that you could play with. Max frame rate & Image sharpening. I don't know how much time nVidia spends on optimizing their drivers after a game was released. 2 years? You will see which games and features got pimped in the driver release notes: https://www.guru3d.com/files-categories ... a-|-7.html
Check them out and determine if there are some benefits for your setup. Otherwise there is little reason to be bothered about latest drivers. Just download the driver you think you need and let 3D fix manager install the 3d driver separately. They've made it so easy for us now.

Re: Performance difference between 425.31 and 452.06

Posted: Mon Oct 12, 2020 6:27 am
by ZeGURU69
Necropants wrote: Sat Oct 10, 2020 4:22 pm Something is funny there. I run 1809 on the lastest driver where it still works and if anything I have better performance.
Sure you used DDU propery in safe mode?
Yes, of course I used DDU in safe mode. Unfortunately, that did not solve the performance problem.

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 23, 2021 5:00 am
by Losti
ZeGURU69 wrote: Mon Sep 28, 2020 10:27 am Hello,

I tried to update my 3Dvision driver on windows 7 from 425.31 to 452.06 via 3Dfix Manager and I noticed a noticeable drop in performance. On games like Divinity 2, Borderland 3 or Vampyr I have at least 30% less performance.

I also have the 452.06 driver on Windows 10 (1809) and it runs like Windows 7. I have an RTX 2080 Ti. Someone has already noticed that the performance drops on the drivers after the 425.31, is there a manipulation to be done to have the same performance ?

Thank you for your answers !
I have exactly the same reults here. Windows 10 20H2, RTX 2080 Ti, I9-9th Gen, 16 GB DDR4 - 4266:
using 425.31 clean installed instead of clean installed 452.06, all settings keept standard after installation:

- 425.31 has EXREMELY increased initial game-loading/save-loading times (up to 200 % approx.). But only for the first start after a new game/driver installation, BUT:
- Days Gone +5-10 % FPS
- Robin Hood - Sherwood Builders: +20 % FPS
- The Ascent: +15 % FPS
- Watch Dogs Legion: +15 % FPS
- Metro Exodus: +10 % FPS

PS: I have used my Driver Changing Tool for the installation of the drivers. For 425.31 i run into the issue, after the installation, that 3D is permanentely enabled, and games are 2D. Its seems to be an issue of my PC-setup. I dont have this issue had before with my 1070 GTX but since 1080 GTX. The solution is to install the 425.31 ON TOP again using normal way to install a driver with the nvidia driver file. Just in case you have the same issue switching to 425.31

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 23, 2021 8:01 am
by harisukro
Make sense those numbers cause Nvidia before every new generation launch - reduces the driver optimization of previous GPUs so imagine even with last driver we would have shitty performance and that's why personally I use 442.74 version.

Thanks Losti for 425.31 benchmark.

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 23, 2021 11:06 am
by Losti
harisukro wrote: Sat Oct 23, 2021 8:01 am Make sense those numbers cause Nvidia before every new generation launch - reduces the driver optimization of previous GPUs so imagine even with last driver we would have shitty performance and that's why personally I use 442.74 version.

Thanks Losti for 425.31 benchmark.
Is there advantage using 442.74 instead of 425.31 for 3d ?? compared to 452.06 ?

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 23, 2021 7:24 pm
by harisukro
Nothing special just more stable, low temperature and less stuttering.

Re: Performance difference between 425.31 and 452.06

Posted: Sat Oct 23, 2021 11:52 pm
by Losti
harisukro wrote: Sat Oct 23, 2021 7:24 pm Nothing special just more stable, low temperature and less stuttering.
Sounds good, ill try it :-)

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 2:49 am
by Chtiblue
Losti wrote: Sat Oct 23, 2021 11:52 pm
harisukro wrote: Sat Oct 23, 2021 7:24 pm Nothing special just more stable, low temperature and less stuttering.
Sounds good, ill try it :-)
Yes please add driver442.74 to your driver changer tool ;)

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 4:43 am
by Losti
Chtiblue wrote: Sun Oct 24, 2021 2:49 am
Losti wrote: Sat Oct 23, 2021 11:52 pm
harisukro wrote: Sat Oct 23, 2021 7:24 pm Nothing special just more stable, low temperature and less stuttering.
Sounds good, ill try it :-)
Yes please add driver442.74 to your driver changer tool ;)
You can simply do this by your own by download it and copy it to the usercustomdriver folder

PS: 442.74 seems to have the same wors performance like 452.06 compared to 425.31

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 3:08 pm
by RAGEdemon
What happens if you use later profiles on older driver, or older profiles on later drivers?

It is conceivable that 3DV optimisations were removed by nVidia in the drivers after 425.31...

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 4:32 pm
by 3DNovice
...

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 9:27 pm
by Losti
RAGEdemon wrote: Sun Oct 24, 2021 3:08 pm What happens if you use later profiles on older driver, or older profiles on later drivers?

It is conceivable that 3DV optimisations were removed by nVidia in the drivers after 425.31...
Nothing, test was of course with latest profiles so they were the same for both drivers.

Re: Performance difference between 425.31 and 452.06

Posted: Sun Oct 24, 2021 9:34 pm
by Losti
3DNovice wrote: Sun Oct 24, 2021 4:32 pm or what about.....

trimming the fat
https://forums.guru3d.com/threads/nvsli ... ty.423072/


More threads/discussions, you can trim the driver so RTX features are removed or included, plus remove whatever else
https://forums.guru3d.com/threads/disco ... on.421390/ (old thread/ended)
https://forums.guru3d.com/threads/496-1 ... on.433367/ (old thread continued)
I have used all the files and packages from the latest driver with 425.31, this do not afflicts performance so i dont think removing someting will also make no difference but ill try. But only to See if 425.31 gets faster. If so, 452.06 will get faster but should be slower than 425.31 annyway.

Re: Performance difference between 425.31 and 452.06

Posted: Mon Oct 25, 2021 6:54 am
by ZeGURU69
Losti wrote: Sat Oct 23, 2021 5:00 am
ZeGURU69 wrote: Mon Sep 28, 2020 10:27 am Hello,

I tried to update my 3Dvision driver on windows 7 from 425.31 to 452.06 via 3Dfix Manager and I noticed a noticeable drop in performance. On games like Divinity 2, Borderland 3 or Vampyr I have at least 30% less performance.

I also have the 452.06 driver on Windows 10 (1809) and it runs like Windows 7. I have an RTX 2080 Ti. Someone has already noticed that the performance drops on the drivers after the 425.31, is there a manipulation to be done to have the same performance ?

Thank you for your answers !
I have exactly the same reults here. Windows 10 20H2, RTX 2080 Ti, I9-9th Gen, 16 GB DDR4 - 4266:
using 425.31 clean installed instead of clean installed 452.06, all settings keept standard after installation:

- 425.31 has EXREMELY increased initial game-loading/save-loading times (up to 200 % approx.). But only for the first start after a new game/driver installation, BUT:
- Days Gone +5-10 % FPS
- Robin Hood - Sherwood Builders: +20 % FPS
- The Ascent: +15 % FPS
- Watch Dogs Legion: +15 % FPS
- Metro Exodus: +10 % FPS

PS: I have used my Driver Changing Tool for the installation of the drivers. For 425.31 i run into the issue, after the installation, that 3D is permanentely enabled, and games are 2D. Its seems to be an issue of my PC-setup. I dont have this issue had before with my 1070 GTX but since 1080 GTX. The solution is to install the 425.31 ON TOP again using normal way to install a driver with the nvidia driver file. Just in case you have the same issue switching to 425.31
Hi Losti, thank you for your observations!

By posting this message I was certain that many people would have made the same observation as me. I was very surprised that this was not the case.

I think that all the riders after 425.31 are affected because I felt a drop in performance as soon as I switched to a rider after 425.31. After that I didn't do a lot of testing in this direction so I'm not sure.

Since then, I have only made games that are not very greedy under win10 452.06 so I was no longer too interested but it reassures me that I am no longer the only one to make this observation!

Re: Performance difference between 425.31 and 452.06

Posted: Tue Oct 26, 2021 7:40 am
by mistersvin21
436.30 is good as well

Re: Performance difference between 425.31 and 452.06

Posted: Tue Oct 26, 2021 11:45 am
by greg02
How is it possible to use anything earlier than 452.XX when windows auto updated and required it? Does Windows 10 20H2 no longer require 452?

Re: Performance difference between 425.31 and 452.06

Posted: Tue Oct 26, 2021 3:08 pm
by RAGEdemon
You can set group policy to never auto update your graphics driver. Most people just disable all driver updates, which isn't a great idea though.

https://www.tenforums.com/tutorials/146 ... river.html

I have made 2 reg files - one for enabling driver change and one for disabling driver change after I change to the driver I want.

Re: Performance difference between 425.31 and 452.06

Posted: Tue Oct 26, 2021 9:22 pm
by greg02
Not sure if Im understanding correctly but the issue wasn't so much the auto update of the driver, but the fact that windows forced me to update and then said 425.XX was no longer allowed. Will that enable me to continue to use 425 even though it is suppoedly unsupported? Thank you!

Re: Performance difference between 425.31 and 452.06

Posted: Tue Oct 26, 2021 9:54 pm
by RAGEdemon
I don't know - I have never had issues installing 425. Maybe things have changed recently?

Re: Performance difference between 425.31 and 452.06

Posted: Wed Oct 27, 2021 3:00 am
by WickedScav
greg02 wrote: Tue Oct 26, 2021 11:45 am How is it possible to use anything earlier than 452.XX when windows auto updated and required it? Does Windows 10 20H2 no longer require 452?
Not sure what you exactly mean, but Windows 10 21H1 obviously still allows you to install basically any driver. Just installed 425.31 on my Win10 21H+ just yesterday without any problems...

Re: Performance difference between 425.31 and 452.06

Posted: Wed Oct 27, 2021 9:01 am
by Losti
greg02 wrote: Tue Oct 26, 2021 11:45 am How is it possible to use anything earlier than 452.XX when windows auto updated and required it? Does Windows 10 20H2 no longer require 452?

Prevention of the driver update by WIN10: USE Sledgehammer tool:

Code: Select all

http://web396.inetseite.de/3dprofiles/Sledgehammer.7z

Re: Performance difference between 425.31 and 452.06

Posted: Wed Oct 27, 2021 9:50 am
by Losti
mistersvin21 wrote: Tue Oct 26, 2021 7:40 am 436.30 is good as well
425.31 is still the best, have tried drivers past 425.31, all that have lower performance compared to 425.31, may be because of the driver hack, not sure.

Re: Performance difference between 425.31 and 452.06

Posted: Wed Oct 27, 2021 10:25 am
by RAGEdemon
Losti wrote: Wed Oct 27, 2021 9:50 am
mistersvin21 wrote: Tue Oct 26, 2021 7:40 am 436.30 is good as well
425.31 is still the best, have tried drivers past 425.31, all that have lower performance compared to 425.31, may be because of the driver hack, not sure.
Thanks for testing Losti - this is the information I have been waiting for.

Also, is it safe to use the latest profile with 425.31 without performance issues?

Re: Performance difference between 425.31 and 452.06

Posted: Wed Oct 27, 2021 10:20 pm
by Losti
RAGEdemon wrote: Wed Oct 27, 2021 10:25 am
Losti wrote: Wed Oct 27, 2021 9:50 am
mistersvin21 wrote: Tue Oct 26, 2021 7:40 am 436.30 is good as well
425.31 is still the best, have tried drivers past 425.31, all that have lower performance compared to 425.31, may be because of the driver hack, not sure.
Thanks for testing Losti - this is the information I have been waiting for.

Also, is it safe to use the latest profile with 425.31 without performance issues?
Sure. I always do so. Of course Nvidia can make games wors in Performance Like Kingdom comes deliverance (long Time ago). In issues with Crashes, strange behavior or Performance i save my game profile delete a game and check without and recover Profile if issue is not solved.

Re: Performance difference between 425.31 and 452.06

Posted: Thu Oct 28, 2021 3:02 am
by greg02
WickedScav wrote: Wed Oct 27, 2021 3:00 am
greg02 wrote: Tue Oct 26, 2021 11:45 am How is it possible to use anything earlier than 452.XX when windows auto updated and required it? Does Windows 10 20H2 no longer require 452?
Not sure what you exactly mean, but Windows 10 21H1 obviously still allows you to install basically any driver. Just installed 425.31 on my Win10 21H+ just yesterday without any problems...
Im still on 19042.1288 and that one won't allow me to install 425.XX. Says this driver is not compatible with this version of windows. Did they roll that back w/ the next update, no longer forcing an updated driver? Was this something I missed?

Re: Performance difference between 425.31 and 452.06

Posted: Thu Oct 28, 2021 3:29 am
by RAGEdemon
You're probably trying to install the DCH drivers over Standard drivers, or vice versa.

Re: Performance difference between 425.31 and 452.06

Posted: Mon Dec 20, 2021 7:35 am
by ZeGURU69
Hello,

I upgraded to Windows 10 21H2 and since unable to use 3D on 425.31 after reboot.

I have blocked the update of the drivers and when I restart, I still have the 425.31 drivers but the 3D does not engage anymore. The only solution (which does not always work) is to reinstall the 425.31 drivers.

Anyone have a solution?