Discussion in 'Videocards - AMD Radeon Drivers Section' started by Komachi, Aug 6, 2018.
Evolution or Revolution
For a link to "previous drivers" scroll down the site...
Cleaner look, easy to navigate ! I love it
Yes, really looks good, i like new design as well. The only thing which is lacking there in comparison to nvidia's website: available previous versions of drivers. To some people this may be useful, currently if you want to download older version driver you need to look at third-party websites like guru3d or other less known/less reputable sources.
if you scroll down , there is a link to "previous drivers".
Yea the new desing looks great, it was needed.
You are right, i missed that. Yes! They finally did it! Still, option should be more exposed as it can be easily missed in its current form.
Better the old drivers be missed IMO, some people are still with 17.1.1 or whatever, if they want to install a driver,
they better find the latest one instant, not a full list of drivers and be confused, and then they go on reddit or random forums and ask ''whats the best driver''...
You know what i mean..
The new download page is pretty straight forward, easy, and gets the job done IMO.
Also the Knowledge Base right under the drivers is very good addition and very visible now.
Adrenalin Edition 18.8.1 Optional - Release Date 8/17/2018 re-upload?
WHQL-Adrenalin-Edition-18.8.1-aug17 is ready to download
@JonasBeckman your Up
Tell Us what's inside of this one.
For those who were curious, this latest build still installs the AMD PCI-E filter driver on Intel motherboards. I did a AMDCleanupUtility and DDU uninstall of the old drivers, made sure the system reverted back to the "PCI Express Root Complex" (pci.sys) driver, and installed these aug17 drivers, and amdkmpfd.sys/pci.sys "Pci Bus" entry returned.
(this was on my old Z97/4790K box)
I have Intel MB and I don't have problem having AMD divers on the PCI bus. I updated them even manually before 18.8.1
"original" drivers are from microsoft, not from Intel, so what's the problem?
These drivers have a problem with gzdoom: enemy/pick up sprites don't show when using opengl renderer. I have no ideia if the issue was already present in other recent amd driver releases. Brutal doom can be quite heavy to run in software mode in an old cpu on modern map packs, so that's not an option. Reverting to a very old gzdoom build 2.0.05 seems to solve it for now, later versions will show the problem.
I have no such issue on a integrated HD 4400 (Haswell laptop).
There are various reasons why someone would like to have an option to download older version of driver. Compatibility, testing, comparisons, having some features which newer driver doesn't have anymore, etc.
There are lots of things written by Microsoft, since it is their OS, lol. This situation is a little different---installing an AMD specific driver for a component that is not created/engineered by AMD, may not sit right with some people.
It is worth noting that I still haven't had a crash or any other weird behavior using it on the Z97 box, but this is a very small sample size for Intel systems.
Thanks, i will install the driver again and upload the next bsod.
Yes i do have Nox App Player installed,( android emulator) maybe is a conflict with this driver.
Deleted Post: Aug 17 Re-release date mentioned above from Aug 3 initial release date.
Aftre your first post i did a ddu clean up and installed the WHQL-Adrenalin-Edition-18.8.1-aug17.
I was gamming and got the bsod.
I was able to get the crash dump. https://www69.zippyshare.com/v/r2FVPaq1/file.html
With a older driver like 15.11 beta i had zero issues.
Thanks again for your time.
Yes, but Microsoft is also not manufacturer of the HW and their drivers are in most cases made "just to work". Like when you have microsoft driver on Intel/nVidia/AMD graphic card. You see picture right?
The best way would be to have true Intel drivers for the PCI but what I know, they don't have them. So maybe that's the reason why AMD makes their own?
"Yes, but Microsoft is also not manufacturer of the HW and their drivers are in most cases made "just to work".
That is a bit different situation---after all, Intel and AMD submit their "out of the box" drivers to MS for WHQL certification, and Microsoft designs their generic "in the box" drivers to work with both platforms universally. If this is also the case with this PCI-E filter driver from AMD getting force installed on both Intel and AMD chipsets, that they designed it for work on both Intel and AMD platforms, then fine.
But given that (a) this has never been the case where it has force installed the filter driver and (b) refused to let you go back to the generic MS driver on an Intel platform (unless you manually edit the inf or revert to an earlier driver), it might be a little concerning to the end user of an Intel platform. Its not like both Nvidia and AMD haven't made these sort of "mistakes" with single driver releases in the past. The thing is, nobody knows if this was intentional or a mistake by AMD currently.
Personally, until my system crashes or exhibits odd behavior, I'm not too concerned, but I can understand why somebody might be.
Seems I missed this update entirely.
From a quick compare though one thing stands out a bit.
These are now part of the signed files when comparing the INF file with these two lines added to it for this section [SignatureAttributes] like so:
And a quick search shows that some anti-cheat software seems to have flagged this .dll previously so I guess it now having a digital signature will resolve that problem.
(RUST back in 2017 for one.)
Not much else other than re-arranging some GPU names and the driver date being August 13th instead of August 3rd.
(A few registry entries were updated for Topaz and Polaris 10 and Polaris 12 entries but it might just be for the GPU names, I'm not too sure what that part of the settings would be used for.)