The fact AMD commissioned the report means nothing. These reports only come about because the company that wants it, pays for it. Guaranteed that the contract states that the results can only be published upon AMD's approval. The publishing of said report is now just a marketing tool. Was solely AMD through my RX480. Now with my 1080 have no more or less issues than had before.

This audit should be taken with a grain of salt

RzrTrek:

This audit should be taken with a grain of salt

Unless end-users want to collaborate on a large scale this is the best data point we've got.

Anyone who thinks Nvidia has driver issues lately may have missed the ATI dark days of pure garbage drivers or no drivers released for a full year or more. People were releasing custom drivers to fix the official ones on a regular basis. There's good reason many still have a bad taste in their mouth when you say AMD drivers. The name may have changed but we still remember. Good to hear they've cleaned up their act.

0blivious:

Anyone who thinks Nvidia has driver issues lately may have missed the ATI dark days of pure garbage drivers or no drivers released for a full year or more. People were releasing custom drivers to fix the official ones on a regular basis. There's good reason many still have a bad taste in their mouth when you say AMD drivers. The name may have changed but we still remember. Good to hear they've cleaned up their act.

Don't worry Pepperidge Farms Remembers.

Since I owned nVidia cards 70% of the time, whenever I had an AMD card it annoyed me that I had to go to control panel/center to change some stuff back to normal settings. Like, no, I don't want you to throttle down my clocks if I am not using the full potential of the card, or no, I don't want to not use multisampling AA, or no, I don't want to use your stupid video profiles, I want them OFF always, no, I don't want the GPU to waste time to do what monitor scaler does for free, and better, and no, I do not want "AMD optimized" settings for textures since they make things loose detail in some games (that I so far noticed) and stuff like that. Oh and good luck with messing with AMD tessellation options, it took me a week to troubleshoot why Black Ops 2 wasn't working at all, until I, by accident, returned the value of that setting to default. But yea, stable as hell, unless they CTD. On the other hand, with nVidia, i just don't install their "nVidia bloatware experience" and I don't even go into control panel, because, no need. No need to install new drivers if the ones I use are working with all the games I play and so on. Hell, I am lying, I do go to nV CP to make a custom resolution (imagine that!) for downsampling. I think AMD made this as some special option, VSR or something that doesn't work on some GPU's of theirs because they are last gen, but the driver could be hacked to circumvent the lie that those cards can't support it? Something like that?! Great stuff from AMD. Oh I also remember Fallout 3, it was a blast on AMD...yea...much drivers, such bugs... but sure, nVidia drivers can CTD at times. Can't remember the last time it happened to me, but I am not exactly the appropriate sample size.

I wonder what NV users say about last 1 or 1,5 years of NV drivers after saying AMD has shxxty drivers. Laughing out loud at the moment.

gx-x:

Since I owned nVidia cards 70% of the time, whenever I had an AMD card it annoyed me that I had to go to control panel/center to change some stuff back to normal settings. Like, no, I don't want you to throttle down my clocks if I am not using the full potential of the card, or no, I don't want to not use multisampling AA, or no, I don't want to use your stupid video profiles, I want them OFF always, no, I don't want the GPU to waste time to do what monitor scaler does for free, and better, and no, I do not want "AMD optimized" settings for textures since they make things loose detail in some games (that I so far noticed) and stuff like that. Oh and good luck with messing with AMD tessellation options, it took me a week to troubleshoot why Black Ops 2 wasn't working at all, until I, by accident, returned the value of that setting to default. But yea, stable as hell, unless they CTD. On the other hand, with nVidia, i just don't install their "nVidia bloatware experience" and I don't even go into control panel, because, no need. No need to install new drivers if the ones I use are working with all the games I play and so on. Hell, I am lying, I do go to nV CP to make a custom resolution (imagine that!) for downsampling. I think AMD made this as some special option, VSR or something that doesn't work on some GPU's of theirs because they are last gen, but the driver could be hacked to circumvent the lie that those cards can't support it? Something like that?! Great stuff from AMD. Oh I also remember Fallout 3, it was a blast on AMD...yea...much drivers, such bugs... but sure, nVidia drivers can CTD at times. Can't remember the last time it happened to me, but I am not exactly the appropriate sample size.

1. Think of the artificial overtessellation levels in Crysis 2 maybe... or Witcher 3 overtessellated HairWorks effects... 2. You are talking about Black Ops 2 and even Fallout 3. We are talking about NV drivers NOWADAYS, not 6 or more years back in time.. 3. NV in the last 1 or 1,5 years released WHQL drivers that simply didn't let Watch Dogs 2 run or instantly crash, or crash with certain browser in many configuration, etc. Pathetic driver quality from a company having much more money for resources than the other.

Denial:

https://www.amd.com/system/files/documents/graphics-driver-quality.pdf Here is the link to the report. Here is what they tested: "We define graphics driver stability as a resistance to Blue Screens (BSODs), application crashes, hangs, and otherwise unexpected behavior in the presence of stressful test vectors over a period of time. We used Microsoft’s Windows HLK as a source of graphics test vectors. Windows HLK is a test framework designed to help hardware manufacturers deliver a quality product to users running Windows operating systems. Within Windows HLK, we used the 64-bit variant of CRASH to stress the Graphics Processing Unit (GPU). CRASH is a GPU stress test tool that spans 4 hours in length and captures test cases covering S3, display resolution changes, display orientation changes, content protection, and rendering. While 4 hours of stress testing is a good indicator of prominent quality issues, it does not suffice in capturing intermittent stability failures or glitches. Therefore, we ran this test back-to-back around the clock for 12 days for each GPU. This accounts for 288 hours of non-stop stress in a test designed to make the GPU driver fail." They didn't test any games - just the HLK "CRASH" test. Which I'd assume being a Microsoft test isn't available on Linux.

They didn't test any 380/480's or they would have shot their test bench with a 12 gauge after 30 mins of testing.

BReal85:

I wonder what NV users say about last 1 or 1,5 years of NV drivers after saying AMD has shxxty drivers. Laughing out loud at the moment. 1. Think of the artificial overtessellation levels in Crysis 2 maybe... or Witcher 3 overtessellated HairWorks effects... 2. You are talking about Black Ops 2 and even Fallout 3. We are talking about NV drivers NOWADAYS, not 6 or more years back in time.. 3. NV in the last 1 or 1,5 years released WHQL drivers that simply didn't let Watch Dogs 2 run or instantly crash, or crash with certain browser in many configuration, etc. Pathetic driver quality from a company having much more money for resources than the other.

Crysis 2 tessellation is a myth - the engine culls unseen geometry when wireframe is off. Witcher 3 HairWorks wasn't overtessellated, the AA was too high and it's a developer set parameter that was later made configurable. Watch Dogs was fixed in 5 days and it only affected a subset of people (i had no issues playing the game during that period) - the firefox thing was an issue but that stuff happens. Overwatch, one of the most popular games, was broken on AMD for half a year. Every company is going to have driver issues that are difficult to fix - throwing money at it doesn't always help solve them.

Who did the audit Lehman Brothers or Goldman Sachs? 🙄

Tough call... AMD did pay for this test. Literally. And I doubt they would go ahead and publish had the results been negative for AMD. So I have trouble calling this an independent testing. But then again this is MS standard GPU stress test, so the results should count for something. Then again the results are heavily skewed by PRO cards failing too frequently, like for example P600 failing 31 out of 72 tests, while 1080 Ti fails only 2 out of 72 tests. How is it possible that p600 is "supported" let alone that this driver is marked as WHQL, if the card fails 50% tests?? Going even further: It doesn't looks like MS testing prerequisites were met, with QA failing to meet the very first prerequisite: At least two test computers for every discrete device family in the INF file that is being certified. https://docs.microsoft.com/en-us/windows-hardware/test/hlk/testref/graphic-adapter-or-chipset-testing-prerequisites I stopped reading "Graphic Adapter or Chipset Testing Prerequisites" because whats the point if QA consultants fail at the 1st prerequisite. I think its a safe bet that there are more testing errors. All this does not mean I would let Nvidia off the hook that easily. I would really like to know what happened here.

I don't know what people doing to their GPUs, but I never had any drivers issue for single GPU. People who run Crossfire/SLI setup are asking for troubles anyway. I think overall drivers stability been increased a lot recently. So having these topics "AMD > Nvidia drivers for gaming" is just plain stupid. We should be past those days. (Better drivers / support for Linux and workloads is totally different talk) "Third party audit" should move on with its life. However, I wouldn't be surprised if AMD marketing team sponsored it. AMD have serious issues with its marketing team and they better stop being desperate with attacking Nvidia and Intel. I understand Zen Epyc attacking Intel Xeon, since it's got something going. But AMD GPU department barking?? Better invest those money into new GPU development.

So funny reading these comments. the amount of passion people have for hardware is hilarious. I wonder if they are the same if it was their printers.

Pimpiklem:

So funny reading these comments. the amount of passion people have for hardware is hilarious. I wonder if they are the same if it was their printers.

EPSON > CANON drivers 😡 😡 CANON FANGAYS can print my ass and kiss it 😡 😡 (oh wait, they can't LUL)

If you watch the video they released you can tell it was all choreographed and staged. The office is fake. the wall sight has been up 5 minutes. It honestly looks like staged nonsense as they pretend to be a active company. It dont work its stinks. this is just bait to get people to respond. I dont mind i find it funny but people are biting.

Pimpiklem:

So funny reading these comments. the amount of passion people have for hardware is hilarious. I wonder if they are the same if it was their printers.

this is like coming to a "Red Rooster And Hen Festival", and complaining MY GOD YOU PPL REALLY LOVE YOUR CHICKEN

sverek:

I don't know what people doing to their GPUs, but I never had any drivers issue for single GPU. People who run Crossfire/SLI setup are asking for troubles anyway. I think overall drivers stability been increased a lot recently. So having these topics "AMD > Nvidia drivers" is just plain stupid. We should be past those days. "Third party audit" should move on with its life. However, I wouldn't be surprised if AMD marketing team sponsored it. AMD have serious issues with its marketing team and they better stop being desperate with attacking Nvidia and Intel. I understand Zen Epyc attacking Intel Xeon, since it's got something going. But AMD GPU department barking?? Better invest those money into new GPU development.

Just because this testing is probably a result of AMD marketing efforts does not mean you should dismiss the results solely based on that. The fact that you don't get to experience any bugs in everyday use does not mean your GFX setup is 100% stable. These tests are (I believe) much more stringent and demanding than just "Play XYZ game" My trouble with this particular testing is that the tester (QA consultant) did not follow MS CRASH testing procedure, and instead they took some liberties in the testing procedure. Furthermore, if they consulted with AMD, at the very least they should have given Nvidia a courtesy call.

Noisiv:

Just because this testing is probably a result of AMD marketing efforts does not mean you should dismiss the results solely based on that. The fact that you don't get to experience any bugs in everyday use does not mean your GFX setup is 100% stable. These tests are (I believe) much more stringent and demanding than just "Play XYZ game" My trouble with this particular testing is that the tester (QA consultant) did not follow MS CRASH testing procedure, and instead they took some liberties in the testing procedure. Furthermore, if they consulted with AMD, at the very least they should have given Nvidia a courtesy call.

reputation and history is the back bone of these kind of test. These people have none. It could of been me doing these test and honestly i would of done a better job at their video. It reminds me of the AMD ryzen video. Its the same thing. Its bait and poorly done bait. Lets point at the screen and pretend i am explaining something about a tea pot render. [youtube=qkaylh0td_E]

Wait... It gets funnier.... This particular test, CRASH, is a part of Microsoft's Hardware Lab Kit Device.Graphics tests. One out of total of 1603 (!!!) tests. https://docs.microsoft.com/en-us/windows-hardware/test/hlk/testref/device-graphics-tests QA consultants singled out this one single test (CRASH) and instead of running CRASH in a standard profile of 240 minutes, they ran it for 12 days straight. So instead of using that testing time of 12 days to run hmmm... lets see... 71 additional test(each running 240 minutes) they chose to ran 1 single test for 12 days straight 🙂

RadioActiveLobster:

Interesting bit left out of the story but included in the source...

Also worth mentioning is that all of the test samples GPUs were supplied and chosen by AMD.